Abstract
The Sahara and Arabian deserts are traditionally thought to be the source of settled dust in the Middle East. Although there have been suggestions of evidence for contributions from local sources, isotopic and clay mineral data have not yet confirmed this. The current study adds the neodymium and strontium isotopic signals as well as the mineralogical composition of the clay fraction to earlier investigations of trapped Holocene eolian sediments in archeological structures carried out in the northern Negev, Israel, and the Petra region, Jordan. Identifying the Holocene and present-day dust source areas and weighing the relative importance of local and remote resources are the objectives of the current study. Illite-smectite (IS) predominates the clay fraction in the Negev, and kaolinite (Ka) in Petra. Clay minerals of both regions form a continuous line of two populations between dust end members of about 75% IS in the Negev and about 75% Ka in Petra. Petra and the Negev have εNd values of −8.8 to −14.1 and −5.1 to −8.6 and 87Sr/86Sr ratios between 0.7082 and 0.7119 and 0.7085 and 0.7134, respectively. The 87Sr/86Sr ranges in the Petra and Negev regions are rather similar, indicating that carbonate rocks from both local and remote sources have had a significant role. It may be inferred that the distinctive εNd values of the Petra eolian sediments must be generated from the local Paleozoic sandstone, even in the absence of information on the isotopic composition of any of the local rocks. Dust from the semi-distant Petra region, the remote Mesopotamia source area, and local sources all have an impact on Negev isotopic signals, but not the Sahara. This is because the significant contribution of coarse-grain local materials masks fine-grain materials from distant sources.
Introduction
Desert dust transportation from the source area to the deposition area through the atmosphere represents current climatic and environmental conditions. Consequently, the accumulation of settled dust from remote and local sources in desert loess sequences can be utilized to reconstruct past paleoenvironments (Crouvi et al., 2008; Grousset et al., 1992; Grousset and Biscaye, 2005; Kubilay et al., 2000; Roberts et al., 2011). Remote materials in the Middle East come from favorable dust source areas (PSAs) in the Sahara and Arabian Deserts. Local sources get their materials from weathered rocks, loose surface sediments, or soils at distances ranging from tens to hundreds of kilometers around sampling points. The grain size distribution, which is affected by both wind speed and distance from the source, has been thought to indicate the remoteness of settled dust sources (Mahowald et al., 2014). Some studies proposed <20 µm for a grain size of remote sources hundreds of kilometers away from the settling location (Kok et al., 2012; Pye and Tsoar, 1987), though later studies also recorded much larger grains for such distances (Beuscher et al., 2020; Ryder et al., 2019). The properties of settled dust and young sediments deposited by settled dust have been utilized for identifying the dust potential source areas (PSAs) and consequently the associated atmospheric and climatic conditions. For many years, such sediments were considered to originate in the Sahara and Arabian deserts (Frumkin and Stein, 2004; Ganor and Foner, 1996; Haliva-Cohen et al., 2012; Smalley et al., 2019; Yaalon, 1997).
Clay minerals and radiogenic isotope traces, primarily of strontium and neodymium, are common proxies for tracing Saharan and Arabian PSAs. Clay mineral investigations of Saharan and Arabian desert dust demonstrated the utility of the illite-to-kaolinite ratio and the presence or absence of palygorskite as valid source proxies (Caquineau et al., 1998; Formenti et al., 2014; Ganor and Foner, 1996; Kandler et al., 2009; Molinaroli, 1996). Accumulating studies on the isotopic signatures of strontium and neodymium in Saharan dust, together with those of lead and hafnium, have been improving the delimiting and characteristics of the African PSAs (Barkley et al., 2022; Guinoiseau et al., 2022; Jewell et al., 2021). However, though the importance of Middle Eastern PSAs has been recognized before (Gherboudj et al., 2017; Yu et al., 2018), their clay minerals have not been reported, whereas their strontium and neodymium isotopic signatures have been reported only recently (Kunkelova et al., 2022).
The strontium and neodymium isotopic values of the insoluble fraction of rocks are regarded as ranging between those of old continental shields with radiogenic 87Sr/86Sr ratios and negative εNd and those of young volcanic rocks with non-radiogenic 87Sr/86Sr ratios and high positive εNd values (Faure, 1977). Detrital sediments, derived from these two isotopic regimes, record the relative share of the source rocks’ isotopic signature (Grousset and Biscaye, 2005). Thus, for example, the Nile’s sediments were found to follow a mixing line between the high 87Sr/86Sr ratios and low εNd values of the White Nile sediments, eroded from Archaean cratons, and the lower 87Sr/86Sr ratios and higher εNd of the Blue Nile and Atbara rivers, which are dominated by eroded Ethiopian volcanic detritus (Fielding et al., 2017). Isotopic data have been the focus of many studies that attempted to locate dust sources in the Middle East. However, data on settled and suspended dust, sensu stricto, are rather limited, as most of these studies used eolian-derived sediments in lakes, caves, streams, soils, and loess to represent the dust composition of their corresponding ages (Ben-Israel et al., 2015; Frumkin and Stein, 2004; Haliva-Cohen et al., 2012; Williams et al., 2022).
Studies on the provenance of Negev loess and young eolian-derived Dead Sea sediments suggested the Nile Delta’s sediments as one of the sources (Ben-Israel et al., 2015; Haliva-Cohen et al., 2012) as well as the Sahara and Arabian deserts (Palchan et al., 2018, 2019; Torfstein et al., 2018). Torfstein et al. (2018) used isotopic signatures in Dead Sea sediments to discriminate between remote and local dust sources. Smaller grains exhibited a Saharan signature (without particular PSA), while bigger grains had a Negev loess signature.
The strontium isotopic signature of suspended dust samples in Israel was found to be 0.7102–0.7190 (Rabi R, 2004, unpublished MSc thesis, The Hebrew University of Jerusalem), more radiogenic than those of settled dust and dust-derived sediments, and corresponding with the remote central Sahara PSAs.
Nevertheless, several studies have emphasized the significance of local sources’ contribution to settled dust and dust-derived sediments in the Middle East, based on field observations and grain size (Crouvi et al., 2008, 2017) or mineralogical and chemical characteristics (Cornille et al., 1990; Lucke et al., 2019b; Sandler, 2013; Sandler et al., 2023c). To better determine the local origins, isotopic data from the insoluble fraction of exposed rocks and surface sediments are required. Though some data on the acid-soluble (carbonate rocks) fraction have been recorded (Frumkin and Stein, 2004; Haliva-Cohen et al., 2012), data on the insoluble fraction, which is required for comparison with the PSA’s data, are limited to loess (Ben-Israel et al., 2015) and certain soil types (Palchan et al., 2018, 2019). Such data for the sedimentary column in Israel and the surrounding countries does not exist, except for the limited data on Cambrian rocks in southern Israel (Ben Dor et al., 2018).
Several studies from Israel have reported on the quantity and composition of clay minerals in settled dust in the Middle East, revealing an amount of 15%–20% and the dominance of smectite (Ganor and Foner, 1996; Sandler, 2013; Singer et al., 2003). The dominant clay mineral smectite has been more precisely defined as a mixture of randomly ordered mixed-layered phases of illite/smectite (IS; Sandler, 2013; Sandler et al., 2023c). Minor associated clay minerals, in descending order, are kaolinite (Ka), illite (I), palygorskite (P), and ordered IS. Similar results were observed with suspended dust (Kalderon-Asael et al., 2009). Loess deposits in Israel and Jordan formed mostly during the last glacial maximum, when dust deposition outpaced erosion (Crouvi et al., 2008; Cordova, 2007). The ensuing change in environmental conditions during the Holocene prevented dust buildup in surface loess but allowed it in artificial structures. In arid climates, eolian sediments trapped by structures can provide insight on the quantity and quality of dust settled between the end of the Pleistocene (~12 ka) and today.
Such sediments have lately been intensively researched in two regions: Sede Boqer in the northern Negev (Israel) and Petra (Jordan), and were characterized as aridic soils (Lucke et al., 2019a, 2019b). Current settled dust and soils from ruins, agricultural terraces, and cisterns were tested for sedimentation rate, texture, and chemical composition. It was found that the chemical compositions of dust and soils in Petra are noticeably different from those of the northern Negev, pointing to the significant contribution of local rocks, sandstone in Petra, and limestone in the Negev. This follow-up study examines the 87Sr/86Sr and εNd isotope signatures, as well as the clay fraction mineralogical composition of selected samples from those earlier studies. Compared to the updated isotopic data on the PSAs of the Sahara with the addition of Arabian PSAs (Kunkelova et al., 2022), and despite the lack of such data for local rocks, it aims to achieve two research goals: 1. evaluating the role of local sources against remote ones, and 2. identifying the source areas of the Holocene and current dust.
Regional setting and sampling sites
The two studied regions, from here on are named Petra and Negev, are located in the southern Levant under an arid climatic regime. Petra is located about 100 km southeast of Sede Boqer across the Dead Sea transform (Figure 1). Each area has its unique geological, geomorphological, and climatic characteristics (see below). However, the synoptic conditions are similar in both areas, as is seen in synoptic maps and wind trajectories that carry dust mainly from the west and southwest (North Africa) or the east (Arabia; Bodenheimer et al., 2019; Dayan et al., 2008; Gherboudj et al., 2017; Kalderon-Asael et al., 2009). Eventually, the eastern location of Petra is supposed to be less influenced by the western trajectories than the northern Negev, though western dust storms are more frequent and have higher ground concentrations than the eastern ones (Bodenheimer et al., 2019).
Map showing the locations of the study areas in Israel and Jordan and the main geological units. Black rectangles mark the studied archeological hilltop soils, black wall symbols resemble terraces, orange asterisks mark the sampled reference sites, and blue hexagons mark the dust collectors in the Petra region. The dust collector in the Negev is located c. 4 km to the south from the sampling sites in Midreshet Ben Gurion.
Both study areas are briefly presented below as a summary of the previously published detailed descriptions (Lucke et al., 2019a, 2019b; Lucke and Bäumler, 2021). The elevation range in Petra is 900–1200 m above sea level, the mean annual precipitation is ~150 mm, and the dominant wind direction is west. Sampling sites included several ruins on Paleozoic sandstones: Jabal Haroun monastery, Jabal Farasha triclinium, Umm Saysaban, and two Early Holocene paleosols below Neolithic sites, besides several archeological structures. A single terrace was located on Turonian dolomitic limestone. A single stream fan was sampled as well. Dust was collected uphill and downhill from Jabal Haroun (Figure 1).
The elevation range in the Negev is 400–500 m above sea level, the mean annual precipitation is 87 mm (1991–2020, Israel Meteorological Service), and the dominant wind direction ranges from the southwest to the northwest. Sampling was performed at two sites: 1. excavated Iron Age ruins and associated agricultural structures on Turonian hard limestone; 2. an agricultural terrace, located on Santonian chalk. Dust was collected at the nearby Sede Boqer College. An additional sample was taken from the southern Negev to present dust composition located on the route of the western trajectories that carry Saharan dust to the Petra area. Because no dust trap could be positioned there, top sediment was sampled as a substitute. The sites and samples list are presented in Table 1.
Location and description of the samples analyzed in the current study.
Methods
Clay fraction (<2 μm) was collected from thin suspensions after carbonate minerals and salts were removed from samples by buffered acetic acid. The residual silicate fraction was disaggregated by a low-intensity ultrasonic treatment for a few minutes. Clay suspensions were pipetted onto glass slides and analyzed after air-drying, glycolation (at least 8 h at 60°C and cooling overnight), and heating for 2 h to 550°C (Moore and Reynolds, 1989). XRD patterns were acquired using a PANalytical X’Pert diffractometer at the Geological Survey of Israel, equipped with CuKα radiation operated at 45 kV and 40 mA. Samples were scanned from 2° to 30° 2θ with a step size of 0.013° 2θ using a PIXcel detector in continuous scanning line (1D) mode. The equivalent time per step was 70 s, resulting in a total measurement time of 11 min per scan. Semi-quantitative clay abundances were estimated in 10% intervals from the relative peak areas of randomly-ordered illite-smectite (IS), kaolinite, illite, palygorskite, and chlorite using known factors (after Sandler, 2013). The estimated errors are ±10%, 20%, 30%, and 50% for amounts of >50%, 20%–50%, 5%–20%, and <5%, respectively.
The Sr and Nd isotopic ratios were measured on the silicate fraction, namely, the whole-rock powder after removal of carbonates and soluble minerals by 0.5 M acetic acid. After three washes, centrifugation, and drying, the silicate fraction was digested with a mixture of fluoric, nitric and hydrochloric acids as well as hydrogen peroxide (5:1:1:0.5). Aliquots suitable for for column separation were prepared. Strontium was separated using Sr spec resin, whereas Nd was separated using TRU followed by LN-spec resins. Sr and Nd isotopes were measured by a Nu Plasma II MC-ICP-MS at the GSI. Sr analyses were corrected to SRM-987 standard (87Sr/86Sr value of 0.71034 ± 0.00026), which was run after every five samples. Nd isotope analyses were corrected to the JNdi standard 143Nd/144Nd value of 0.512115 ± 7 (Tanaka et al., 2000) and run every five samples. Nd isotopic ratios are calculated as εNd=[[Nd143/Nd144 (meas.)/143Nd/144Nd (CHUR)]−1] × 104, where the CHUR value is 0.512638.
Statistical significance (p value) was calculated by the T.TEST function in Excel.
Results and interpretation
Clay minerals
The clay mineralogical composition of the clay fraction is presented in Table 2. IS and kaolinite together consist of more than ~80% of the clay fraction in sediments near Petra and more than ~60% in the Negev. Illite in Petra is around 5%, or less, except for one sample of 5%–10%, whereas in the Negev, it is mostly around 5% and up to ~15%. Palygorskite of <5 to ~15% amounts occur in Petra in several samples and the Negev in most samples. Ordered IS, chlorite, goethite, and quartz may occur in trace amounts of a few percent each.
The mineralogical composition of the clay fraction (in %).
IS: illite-smectite; Ka: kaolinite; I: illite; IS or: ordered IS; P: palygorskite; Ch: chlorite; Qtz: quartz; Goet: goethite.
Samples from both areas form a continuous line (R2 = 0.93, p < 0.05) of two populations between dust end members: ~75% IS (+I) in the Negev and ~75% Ka in Petra (Figure 2). The kaolinite-rich local sandstone (Amireh et al., 1994) is the only source of the Petra kaolinite-rich end member, as evidenced by the high amount of sand fraction in the collected dust (Lucke et al., 2019a). The source of the Negev IS-rich end member cannot be specified, as IS is dominant in both local rocks and remote suspended dust (Kalderon-Asael et al., 2009). Both locations’ terraces are nourished by local erosional materials and dust, and the clay composition of Petra’s dolomitic limestone coincides with that of the Negev population.
The dominant clay minerals, IS (+I) in the Negev (blue symbols) and kaolinite in Petra (red symbols), form a linear correlation (R2 = 0.93; p < 0.05) between two endmembers: Petra dust (two overlapping samples) and the Negev dust. The clay minerals display a distinct separation of the two regions.
At Site 33, clay minerals diversity between the bottom, interpreted as paleosol, or clay-rich sediment of a past water-collecting basin, and the top samples must reflect discontinuities in its sedimentary history. At the lowest layer, remote dust accumulation was the primary contributor. Later, local material with kaolinite, quartz (reflected by SiO₂ concentration), and a texture dominated by fine sand was formed.
The uppermost sample is enriched by carbonates and related IS, indicating additions from carbonate outcrops. Some of the clay-associated chemical parameters display a gradual increase downward, along with the amount of the clay fraction, which might indicate clay illuviation (Supplemental Appendix). Sandy soils along the central coastal plain of Israel, which exhibit clay illuviation, are generally slightly acidic but occasionally alkaline (Sandler et al., 2023b). Moreover, sandy soils under semi-arid and arid climates with pH above 7.5 exhibit clay illuviation (Koyumdjisky et al., 1988, Selected soils profiles from Israel. Volcani Center, Bet Dagan, in Hebrew. Examples on pages 25, 60, 64, 66, and 196). At Site 2 (the Monastery Garden), however, a mutual gradual change with depth occurs in both clay mineralogy and in many of the chemical parameters (Supplemental Appendix). This gradualism might suggest incipient clay pedogenesis.
The ruin soils in the Negev region have a consistent composition of ~45%–60% IS and 20%–30% kaolinite, suggesting a shared source. The terrace on chalk (HaRo’a Farm) shows a mutual gradual increase of the clay fraction and some of the chemical parameters with the depth (Supplemental Appendix). In contrast, the ruins’ soil (HH CW R) displays a mutual gradual decrease in depth of the amount of the clay fraction, with a corresponding decrease of some chemical parameters. However, some other chemical parameters increase with depth. The underlying Late Pleistocene paleosol is not a part of these trends, indicating a different parent material. The Negev dust, including the similar southern Negev sample, has lower kaolinite and higher IS amounts, on average, than the soils. This difference suggests that the current dust is less impacted by eastern sources than the Petra region.
Nd-Sr isotope ratios
The neodymium and strontium isotopic signals of the silicate fraction are presented in Table 3. The εNd values of the soils and dust of Petra (Figure 3) are in a range of -8.7 to −14.1, lower than the values of soils and dust of the Negev, and indicating different sources. The soil on dolomitic limestone has the highest values, close to some of the Negev values, indicating substantial contribution from the bedrock. The 87Sr/86Sr ratios in Petra are in the range of 0.7081–0.7119, similar, in average, to the ratios in the Negev. The values of the soil on dolomitic limestone are the lowest, indicating again contribution from the bedrock.
The isotopic composition of strontium and neodymium.
Standard error for all samples < 0.00002.
Standard error for all samples < 0.00001.
A plot of εNd against 87Sr/86Sr showing: 1. a relatively narrow 87Sr/86Sr range in Petra, except the samples of a terrace on dolomitic limestone (marked by circle), which are similar to the Negev samples, and 2. a relatively narrow εNd range in the Negev.
The εNd values of the soils and dust of the Negev area are in a narrow range of −5.1 to −8.6, whereas the 87Sr/86Sr ratios are in the range of 0.7085–0.7134 (Table 3; Figure 3). The Petra and Negev dust share, except for the rain-washed sample, a narrow range of 87Sr/86Sr ratios. Though the 87Sr/86Sr ratios of the Negev and Petra soils are similar, on average, most Petra ratios are higher than the ratios of the Negev (Figure 3).
The εNd values were examined against clay minerals, grain size, and chemical parameters to identify factors controlling their distribution.
No relation was established with the clay type, but a positive correlation (R2 = 0.45, p < 0.05) with the amount of clay fraction was established for the Petra samples, excluding a paleosol outlier (Figure 4a). The Negev samples do not show such a correlation, suggesting an εNd contribution by a single main source or several similar sources. The εNd against fine silt plot displays, however, a positive correlation for both populations (R² = 0.53, p < 0.05), excluding a dust outlier of high fine silt content in a rain-wash dust sample (Figure 4b).
Plots of εNd against: (a) clay fraction; (b) fine silt fraction; (c) iron oxide; (d) Fe(d/t); and (e) calcium oxide. The Petra dolomitic limestone samples are encircled.
Both the Petra and Negev populations, excluding two outliers, display a positive correlation (R2 = 0.67, p < 0.05) of the εNd against iron oxide plot (Figure 4c). This correlation is assumed to be result of neodymium-bearing primary minerals, as monazite, occurring in the fine silt fraction along with iron minerals. Petra εNd values are positively correlated with Zr and Th (R² = 0.44 and 0.52, respectively, both p < 0.05), not shown here, suggesting association with primary minerals such as zircon and monazite, respectively.
The parameter Fe(d/t) presents the ratio of dissolved to total iron and is used to estimate the share of pedogenic iron minerals in soils. Figure 4d displays a negative correlation of εNd to most Fe(d/t) values (continuous arrow). Since the aridic soils studied here could hardly support chemical pedogenesis (see below), the Fe(d/t) values do not reflect current pedogenesis but preserve the source values, as has been suggested earlier (Lucke et al., 2019a). Thus, the Petra higher values are in accord with intense Paleozoic pedogenic conditions, evidenced also by the dominance of kaolinite. Figure 4d also suggests the mixing of two source populations and a possible third minor source, shown by the dashed arrow.
Keeping in mind that the isotopic values were measured on the silicate fraction and the chemical parameters on the bulk samples, the εNd against CaO diagram displays a distinct separation of the Petra and the Negev populations (Figure 4e). Low CaO concentrations and low εNd values characterize most of the Petra samples, and vice versa in the Negev. Thus, the concentration of carbonate minerals does not affect its neodymium isotopic signature in each of the populations.
The 87Sr/86Sr ratios were examined against all available parameters as well, but a correlation was observed only with two of them. A negative correlation with Zr was observed for all samples (R² = 0.55, p < 0.05), excluding the dolomitic limestone samples and another Negev outlier (Figure 5a). As zircon is not supposed to contain strontium, it seems that other minerals that accompany it, mainly feldspars, create this correlation. The chemical parameters cannot support this, since carbonates and clays, with high calcium and potassium, mask the involvement of feldspars. The 87Sr/86Sr against Fe(d/t) plot displays a positive correlation for the two populations (R² = 0.56, p < 0.05), excluding the dolomitic limestone samples and another Negev outlier one (Figure 5b). The separation of the two populations suggests mixing between two main sources, supporting the suggestion that the Fe(d/t) values are related to the bedrock and not to current soil processes.
Plots of 87Sr/86Sr against: (a) zircon. A negative correlation with Zr was observed for all samples (R² = 0.55, p < 0.05), excluding the dolomitic limestone samples (delineated), and another Negev outlier; (b Fe(d/t). A positive correlation for soil samples (R² = 0.56, p < 0.05), excluding the dolomitic limestone samples (delineated), and another Negev outlier.
Discussion
Local dust sources
The ruin soils in both research regions were found to be mixtures of roughly equal contributions from local and distant sources, according to statistical analyses of chemical data and grain size conducted in the former studies (Lucke et al., 2019a; Lucke et al., 2019b). The clay minerals distribution, determined in the current study, clearly supports the former studies. It displays an obvious distinction between the two regions that reflect the local lithology, but also some contribution of each region, or similar regions, to the other (Figure 2). The 100 km distance between the two regions is an intermediate position between what is generally considered as local and remote sources (Kok et al., 2012; Pye and Tsoar, 1987), but here it would be considered as semi-remote.
The contribution of local sources to the dust and Holocene archeological soils in the Petra region is well pronounced by both the high kaolinite content, dominating the clay fraction in the Paleozoic sedimentary rocks, and the very low εNd values (see below). Moreover, a local distinction is also well observed within the Petra region itself, where the terrace soil on dolomitic limestone is similar in most parameters to the Negev soils and not to those local sites. As stated above (Regional setting and sampling site), the same trajectories are affecting both regions, with the significance of west-direction dust storms from Sahara. Accordingly, the dominance of remote sources in the settled dust in both areas would suggest similar compositions. This assumption can apparently be supported by chemical composition results. Settled dust, collected in ten locations in Jordan under the same trajectory type (Khamseen/Sharav; Abed et al., 2009), had an average chemical composition that was found to be similar to dust accumulated within a pit in the Northern Negev (Sandler et al., 2023c). However, this is not supported in the current study by the isotopic compositions.
To the best of our knowledge, no data on the isotopic signatures of the Paleozoic rocks in the Petra region in particular, or in Jordan, in general, is available. The nearest such data exists for Paleozoic (Cambrian) rocks in south Israel, where separated K-feldspar and clay fractions were found to have εNd values of −8 to −14, which could also represent the local Petra sandstones source, and an 87Sr/86Sr ratio range of 0.704–0.746, which is too wide to be related to any documented source (see below). Consequently, the Paleozoic sandstones at, and near Petra should have an estimated 87Sr/86Sr ratio range of around 0.709–0.712.
At the same time, all current dust samples converge into a narrow range of 87Sr/86Sr ratios, 0.7096–0.7110, despite the different neodymium signal ranges. Accordingly, several sources, local or remote, have similar strontium isotopic composition for both regions. The somewhat higher values of the Holocene soil ratios in Petra indicate contribution from the local sandstone, where the somewhat lower ratios in the Negev soils suggest contribution from the local lithology, which is Cretaceous limestone and chalk.
To the best of our knowledge, no published strontium and neodymium isotopic ratios of the silicate fraction of carbonate rocks in Israel and Jordan are available. Nevertheless, their strontium isotopic ratio can be estimated by the current results to be as 0.7080–0.7085, based on the strontium ratios of high carbonate contribution, as the terrace on dolomitic limestone in Petra. This estimated range resembles the 0.7081–0.7082 ratios of the Valley Loess, which is a surface cover found west of the Dead Sea on chalk bedrock (Palchan et al., 2018).
To decipher the potential local sources of dust in the Holocene archeological soils, the isotopic signals of the current study were plotted along with all relevant data published (Figure 6). Dust from central Israel (Palchan et al., 2018) was found to have εNd values similar to the current Negev samples but higher than the Petra values. Its 87Sr/86Sr range, however, is much wider, extending beyond the ratios of the Petra region and the two exceptionally high values from the Negev, and within the range of 0.7102–0.7190 of suspended dust in central Israel (Rabi R, 2004, unpublished MSc Thesis, The Hebrew University of Jerusalem). The finer texture of suspended dust reflects the isotopic signals from remote sources, including Central Africa (see below). Accordingly, settled dust of relatively high 87Sr/86Sr values is inferred to contain more fine particles from remote sources. Thus, the high ratio of the Negev cistern fill may reflect grain size winnowing and accumulation of finer material, which could be connected with the use of settling tanks along channels carrying water to the cisterns in order to reduce the amount of sediment in the water. Enrichment of the clay fraction by high 87Sr/86Sr ratios has been considered a systematic phenomenon (Garçon et al., 2014; Goldstein and Jacobsen, 1988) and was also observed for the Nile delta cone sediments (Fielding et al., 2018). However, here, the combined population of all dust samples presented in Figure 6 does not follow a mixing line as that of the Nile.
A plot of εNd against 87Sr/86Sr of the current results and associated local material from published data. The gray field encompasses all Negev pristine loess data (Ben-Israel et al., 2015). Additional data: central Israel dust and Terra Rossa (Palchan et al., 2018), Jordan surface sediments (Kunkelova et al., 2022), and secondary loess (Haliva-Cohen et al., 2012).
The Negev loess of the Late Pleistocene age has been suggested as a source for dust (Crouvi et al., 2018; Sandler et al., 2023b). Primary (top hill) Negev loess was analyzed for two size groups of <20 and >20 µm (Ben-Israel et al., 2015), both are included as one field in Figure 6, besides a few samples of secondary loess, which is lowlands loess, mainly transported by streams (Haliva-Cohen et al., 2012). The loess field is larger than the current Negev results, except for the two outliers, reflecting the much longer accumulation time, which included temporal changes in dust directions and remote sources (Ben-Israel et al., 2015). However, most of Petra’s and some central Israel dust samples are lying outside the pristine loess field.
Another reference group is from central Jordan (Kunkelova et al., 2022). Several surface sediments (<37 µm), were found to have 87Sr/86Sr ratios higher than most Petra, the Negev, and most pristine loess samples. Unexpectedly, the narrow neodymium range of −6.7 to −7.8 is much higher than that of the Petra samples but is similar to that of most Negev samples. These isotopic signals suggest either that the surface sediments have more fine grain size particles of remote sources than the other groups in Figure 6, or that they are contributed by another unknown local source or both.
Dust, local and remote, is also the parent material of Terra Rossa soils in Israel, (Ben-Asher et al., 2019; Crouvi et al., 2008; Sandler et al., 2023c) so these soils should share similar isotopic signals with those of the Holocene soils and current dust. Unexpectedly, this is not the case, especially concerning the 87Sr/86Sr ratios, which are mostly higher than all the Holocene and Late Pleistocene Negev samples and most Petra samples, in accord with the dominance of the clay fraction in this soil type (Sandler et al., 2023c). In addition, Terra Rossa soils are formed under the Mediterranean climate regime and continuously respond to the local environmental conditions. Consequently, newly formed or transformed, clay minerals may incorporate soil solution of higher strontium ratios from dissolving feldspars (Sandler et al., 2023a, 2023b).
The Terra Rossa samples have a narrow neodymium range of −5.1 to −7.6, which corresponds to the higher part of all Negev samples but is even higher than that of the central Jordan surface sediments. Therefore, Nd isotopes, which have been considered not to undergo major fractionation during weathering and transport (Feng et al., 2009; Grousset and Biscaye, 2005; Jung et al., 2004), follow the strontium signals and display higher values in clayey soils than in dust and desert soils. Dependence of neodymium signals on grain size has already been observed in the Negev pristine loess, where the <20 had higher values than the >20 µm fraction (Fig. 4 in Ben-Israel et al., 2015), and recently recorded in stream sediments in West Africa, with higher signals in the clay than in the sand fraction (Bayon et al., 2024).
In summary, the following key points emerge when compared to published data on dust and related sediment groups. 1. The Petra dust and soils show the dominance of local sources, as they differ not only from the Negev and central Israel groups, but also from the surface sediments of central Jordan. 2. The Negev dust and soils group is located at the higher εNd and lower 87Sr/86Sr part of the late Pleistocene loess group, indicating temporal changes in dust dynamics. 3. The strontium and neodymium signals are impacted by the grain size distribution.
In situ compositional changes
The gradual variations in soil profiles shown above (Results) point to pedogenic processes, which could be resolved for one of them. The monastery garden (JH site) in the Petra region showed no visible traces of pedogenesis or subsequent carbonate precipitation. However, the steady decrease in IS is accompanied by an increase in the amount of clay fraction as well as by an increase of Al2O3, Fe2O3 and K2O concentrations, which are associated with clay minerals. The concomitant drop in CaO and MgO implies modest carbonate dissolution, as well as a slight rise in absolute levels of dithionite-extractable (pedogenic) iron. These findings indicate that irrigation and manuring may have enhanced pedogenesis in comparison to other profiles. This is supported by a major change in either of the isotopic signals. The 87Sr/86Sr ratio is 0.7115 in the upper sample, like most Petra samples, but is 0.7095 in the lowest sample (Table 3), indicating either an older different parent material or the dissolution of some calcium carbonate of low 87Sr/86Sr ratio into the soil solution. The 87Sr/86Sr ratio of the carbonate fraction of dust and soils in Israel is mostly 0.7080–0.7087 (Palchan et al., 2019). Carbonate mineral dissolution is known to have a greater impact than silicate minerals on strontium ratios in the environment (Jacobson and Blum, 2000; Oliva et al., 2004), and therefore, even a small quantity could affect the pedogenically formed IS. Evidently, if such results occur where the archeological context is not known, they could suggest cultivation. In this context, resolved cultivated soils cannot be utilized for climate reconstruction.
The timing of pedogenic iron minerals formation, represented by the Fe(d/t) values, is cleared up by the relation to the isotopic signals and clay minerals. The Petra Fe(d/t) values are higher than in the Negev, and are associated with much higher kaolinite content, an indicator for significant leaching (Table 2). As such, they reflect the bedrock contribution and are the result of the Paleozoic pedogenesis. This is consistent with previous research that suggested the magnetic signal in ruin soils was connected with dust rather than in situ pedogenesis (Lucke et al., 2019a, 2019b). Furthermore, the clay minerals (Figure 2) and isotopic signals (Figures 4 and 5) strongly imply the mixing of two main sources for each region, albeit the vast spread of the results cannot rule out a slight impact of pedogenesis.
Remote dust sources
The relation of strontium to neodymium isotopic signals (Figure 3) does not follow the traditional mixing line between basaltic materials and granitic materials. In the Middle East, this relationship has been observed for the Nile Delta and fan (Fielding et al., 2017; Revel et al., 2015), and it has also been suggested for surface sediments in Israel (Haliva-Cohen et al., 2012). Here (Figure 3), it is assumed that the source end-members are different. One remote end member source has a low 87Sr/86Sr ratio of around 0.7080 and a relatively high corresponding εNd value of around −3. A second end member is around 0.711 and −15, whereas a possible third end member is of a contradictory trend of a high 87Sr/86Sr ratio of around 0.714 and a high εNd value of around −5.
As stated above (Regional setting and sampling site), the same dust storm trajectories are affecting both regions, with west-direction dust storms from Sahara being particularly significant. As a result, the dominance of remote sources in settled dust in both regions would suggest similar compositions. This assumption can apparently be supported by chemical composition results. Settled dust, collected under the same trajectory type (Khamseen/Sharav) in 10 locations in Jordan (Abed et al., 2009) had an average chemical composition, which was found similar to that of dust accumulated within a pit in the Northern Negev (Sandler et al., 2023c). However, this is not supported by the isotopic compositions of the current study.
Characterization of the PSAs by strontium and neodymium isotopic signals has been improving in the last decade due to the addition of more isotopic data and dust emission potentials. The current results are compared to the PSAs characterizations of Kunkelova et al., (2022), who added characterization of Middle East PSAs. Based on satellite-derived maps of dust source activation frequency, the weighted means (±1 sd) of 87Sr/86Sr and εNd of unconsolidated surface sediments in PSAs were calculated to characterize each PSA. The relevant PSAs are presented in Figure 7. In addition to PSAs data, two Nile data sets were plotted as well (Figure 7) to evaluate Nile-derived sediments as parent materials for dust and loess in the Negev.
A plot of εNd against 87Sr/86Sr of the current results, of the Nile, and of relevant PSAs. The Nile data are from Fielding et al., (2017) and Revel et al., (2015). The PSAs data are from Kunkelova et al., (2022). ES: eastern Sahara; Mes: Mesopotamia; A: Arabia; CS: central Sahara. The Negev results are associated with the Mesopotamia and Arabia PSAs, whereas the Petra results are not associated with any of the PSAs.
The position of the current results of Petra and the Negev regions relative to the relevant PSAs’ fields excludes the West Africa field as a reference for comparison. Moreover, the East and Central Sahara PSAs can be excluded as main sources of remote dust. The current 87Sr/86Sr ratios are too low relative to those of Central Sahara and too high relative to Eastern Sahara ratios, and the εNd values are too low to the East Sahara values. Whereas the Negev εNd could have been conceived as falling on a mixture line between the Central and Eastern Sahara sources, the corresponding narrow 87Sr/86Sr range negates this option.
The very high strontium ratios in two Negev samples, one dust and one cistern fill, indicate that finer material was occasionally added from a remote source, or that such fine material was selectively deposited in the cistern. This source could be the above-suggested third end member (Figure 3) with the isotopic signals of 0.715 and −5. This source could be derived from Central Sahara source with some additional contributions from Arabia. The Negev isotopic signals are rather close and partially overlap the Arabia field. However, the apparent missing impact of Eastern and Central Sahara, known to contribute much of the dust arriving at the Negev and central Israel from the west and southwest (Bodenheimer et al., 2019; Dayan et al., 2008), suggests again that local sources are dominating. The Negev isotopic signals overlap a narrow part of the Nile’s trend line, having large ranges for both neodymium and strontium, as it blends grains from the basaltic Ethiopia tributaries and the White Nile granitoid territories (Fielding et al., 2017; Revel et al., 2015). As the dunes of northeastern Sinai and northwestern Negev were derived from past exposed Nile Delta sediments, it has been proposed that the Negev loess is an erosional derivation of the adjacent dunes (Crouvi et al., 2008; Muhs et al., 2013). This has also been suggested to be evidenced by the isotopic signals of the fine fraction (<20 mm) of the Negev loess, along with some temporal additions from Arabia (Ben-Israel et al., 2015). However, the isotopic signals of the dunes have not been determined. Moreover, the assumption that the Nile and the Sahara sources have the same isotopic signals, as suggested by Ben-Israel et al. (2015), is not supported by the updated PSAs data presented by Kunkelova et al. (2022). In addition, the dunes contain a calcite and clay fraction of up to a few percent (Muhs et al., 2013; Roskin et al., 2011). In contrast, their supposed derivative, the Negev pristine loess, contains, on average, 21% calcite and 17% clay minerals (Crouvi et al., 2008), which had to originate from other sources. Therefore, the dunes cannot be considered a major local source. Considering the remote sources, the Mesopotamia and Arabia fields are major contributors for the current dust and Holocene archeological soils.
Nevertheless, an additional unidentified source(s) of a similar strontium range but a lower neodymium range is still missing. This could be the Petra region dust, as supported by the above kaolinite data (Figure 2). It has already been shown above that clay fraction has higher strontium ratios than their corresponding coarser fractions, that the silicate fraction of carbonates has low estimated strontium ratios, and that both contradictory trends might affect the Negev signals. Accordingly, it is suggested that the Negev isotopic signals result from the mixing of local carbonate rocks and two remote sources, Mesopotamia and Petra dust.
In summary, dust and Holocene archeological soils in Petra and the Negev contain a significant portion of local material, derived from weathering and erosion of surrounding rocks. As a result, the local contribution largely masks the portion contributed by remote sources. The bulk compositions of settled dust, including the <20 µm portion, are supposed to be insufficient for identifying remote dust sources. After taking this into account, sediments that are formed from dust, including loess, stream sediments, and sea sediments, which eventually contain contributions from local sources that are greater than dust, are even inferior in this regard. Analysis of suspended dust, as well as separated fine silt and clay fractions in settled dust and dust-derived sediments, is advised in order to better identify dust source areas and to gain a better understanding of past climatic changes from dust-derived sediments. Future determination of the silicate fraction of the local rocks is a prerequisite to better identify the local sources.
Conclusions
The clay mineralogy and εNd values of current settled dust and Holocene archeological soils in the Petra region differ from those of the Negev region, despite being affected by the same dust storm trajectories.
The 87Sr/86Sr ranges of the Petra and Negev regions are rather similar, indicating contribution from carbonate rocks of both local and remote sources. As the silicate fraction of these rocks has not yet been determined, it is estimated here to be in the range 0.7080–0.7085.
The 87Sr/86Sr ratios of two exceptional Negev samples are comparable to those of suspended and some central Israel settled dust, suggesting that they have a fine texture and may have come from the Central Sahara PSA.
The Petra isotopic signals are unique and cannot be related to any of the PSAs. Apparently, they reflect the local sandstone composition, which its εNd and 87Sr/86Sr values have not yet been determined.
The Negev isotopic signals are mixtures of local sources, the remote Mesopotamia PSA, and the semi-remote Petra region.
Pedogenic processes in the aridic archeological soils studied, if they occurred, are rather undetectable, except, at one profile that experienced irrigation and manuring.
Due to the significant local influence on the 87Sr/86Sr and εNd signals of settled dust, it is challenging to identify the remote sources in dust-related deposits such as loess, soils, and lake sediments.
For better identifying settled dust sources, it is necessary to determine the isotopic composition of several size fractions and to construct an isotopic database of the local rocks.
Supplemental Material
sj-xlsx-1-hol-10.1177_09596836251387247 – Supplemental material for Settling dust sources: Clay minerals, 87Sr/86Sr and εNd in Petra (Jordan) and the northern Negev (Israel) regions point to the importance of local sources
Supplemental material, sj-xlsx-1-hol-10.1177_09596836251387247 for Settling dust sources: Clay minerals, 87Sr/86Sr and εNd in Petra (Jordan) and the northern Negev (Israel) regions point to the importance of local sources by Amir Sandler and Bernhard Lucke in The Holocene
Footnotes
Acknowledgements
We thank Tzahi Golan and Keren Weiss (GSI) for performing the isotopic analyses, and Navot Morag and Olga Berlin (GSI) for performing the X-Ray diffraction runs. We thanks Editor Yama Dixit and two reviewers for their helpful comments.
Author contributions
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the German Research Foundation (DFG) under Grant Agreement No LU 1552/2-1.
Data availability statement
All data generated or analyzed during this study are included in this published article [and its supplementary information files].
Supplemental material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
