Lanthanides Revealing Anthropogenic Impact within a Stratigraphic Sequence

Difficulties to differentiate between anthropogenic and natural processes in the formation of archaeological deposits are crucial for a correct interpretation not only of the actions involved in the development of archaeological sites, but also of their occupationabandonment dynamics and the understanding of their spatial behaviors and relationship with the environment. We have carried out lanthanides (rare earth elements “REE”) analysis to distinguish anthropogenic from natural stratigraphic units in sediments using the advantage of the high sensibility, precision, and accuracy of ICP-MS measurements. In the Neolithic site of Mas d’Is (Alacant, Spain), we have applied REE analysis in a huge stratigraphic sequence called Pit 6, which was known to contain a large anthropogenic component. Randomly collected soil samples were sequentially taken in order to identify anthropogenic soil formations and to prove the proposed method blind testing has been used. In the specific case of Mas d’Is excavation a recurring question is whether paleosols are at the origin of the human occupation of the sites or it was the occupation of this areas which triggered the paleosols development. Our purpose was to distinguish the degree of human contribution to paleosols formation between samples sequentially taken at few centimeters of distances in a giant stratigraphic sequence (Pit 6) employing REE analysis.


Introduction
Neolithic site of Mas d'Is (Alacant, Spain) and its hinterland (Figure 1) has led us to face some problems about the interaction of anthropogenic and natural processes [1].At Mas d'Is we have discovered huge Neolithic stratigraphic sequences >5 meters that are not at all common in western Mediterranean for an open-air site [2] or elsewhere for the filling of some previously excavated features.New methodological proposals that can contribute to a better understanding of some sedimentological problems in this archaeological site have been tested.We have successfully applied lanthanides (rare earth elements "REE") analysis to distinguish anthropogenic from natural stratigraphic units in predefined classes [1].The potential of REE to distinguish anthropogenic versus natural sediments was useful in Mas d'Is case, where other wellknown methodologies (sedimentology, microsedimentology, and petrography) fail in discriminating between human and natural soils [1].In this study we have applied REE analysis in one of those huge stratigraphic sequences called Pit 6.
Randomly collected soil samples were sequentially taken in order to identify anthropogenic soil formations and to prove the proposed method blind testing has been used.
Recently, many studies have concentrated their attention in different environments to know how human activity is recorded in soil formation.Major and trace elements chemical analysis has a long standing application in archaeology [3][4][5][6].Multielemental soil analysis has been employed as a tool of geochemical prospecting [7][8][9][10] and to support interpretation of past workplaces and activity spaces around and within archaeological structures [11,12].In the last three decades REE soil analysis has started to be part of archaeometric studies.It is the strong partition of rear earth elements into the mineral particulate phase, their coherent behavior during weathering, erosion and fluvial transportation, and their high resistance to chemical mobilization [13][14][15][16] which underlie the use of REE in sediments as provenance indicators in geochemical studies for some time now [17][18][19][20][21][22].Due to their peculiar characteristics, there have been several attempts to apply REE soil analysis to different archaeological scenarios [23][24][25][26].
At the Mas d'Is, to test our methodological approach we have analyzed 15 soil samples.REE were measured using inductively coupled plasma mass spectrometry (ICP-MS).The multielement capabilities, high sensibility, precision, and accuracy of ICP-MS have been exploited to evaluate variations on REE between our samples.In the specific case of Mas d'Is excavation a recurring question is whether paleosols are at the origin of the human occupation of the sites or it was the occupation of this areas which triggered the paleosols development.Our purpose was to distinguish the degree of human contribution to paleosols formation between samples sequentially taken at few centimeters of distances in a giant stratigraphic sequence (Pit 6) employing REE analysis.

Material and Method
2.1.Geographical and Archaeological Background.The samples were collected on Mas d'Is archaeological site located in the mountainous inland area of the northernmost Alacant province (Figure 1), within Penàguila headwaters, one of the main Serpis tributaries.The area where the site is located is an old platform where heavy erosion processes have led to a series of deep ravines.Holocene landscape development appears to have progressed from a period of stability to slope denudation with aggradation (stream infilling) followed by rapid incision which initiated sometime near the time of occupation-in fact.Some structures of Mas d'Is are eroded by ravines, which confirms that some incision process occurs after early Neolithic occupation (ca.5600-4800 cal.BC).The result is a series of deep canyons cut by interfluves receiving the distinctive name of Les Puntes, one of which houses the Neolithic site of Mas d'Is.The current profile is the result of ongoing agricultural changes that ultimately have levelled the previous agricultural terracing topography masking and destroying much of the close-to-surface archaeological record.
The sedimentological environment at Penàguila valley can be characterized as follows: on top of the bedrock, very homogenous beds of Miocene marls and almost 1000 meters thick a set of sediments ranging from weathered marls to alluvial deposits.Those alluvial deposits are framed by the development of important paleosols predating the Neolithic occupation of the site.Around ca.5600 cal.BC, the first farmers appear at Mas d'Is, on the top of one of these paleosols.
Archaeological investigations, beginning at the end of 19th century, have identified an important number of human sites dated to the Early and Middle Holocene.More recently, extensive surveys and excavations of Neolithic villages have provided more detailed information about human settlements in the Alcoi valley.Currently, we have documented more than 100 Neolithic open-air sites that almost without exception are related to paleosols.
The excavated area at Mas d'Is provides different kinds of structures and archaeological materials that chronologically span between the very first Neolithic in the region (ca. ) and present times.The space occupied by the Neolithic site has been an agricultural plot since the Early Neolithic.In the sixth millennium cal BC the site could be divided in two main areas: the village, where the domestic features (houses, hearths, and mills) are found [27] and the area of the ditches, over 150 m east.Some ditches have an important sequence of infilling sediments, covering more than one millennium (from ca.5600-4000 cal.BC).On the top, a disturbed layer of Bronze Age ends the archaeological occupation of this area [1].Pit 6 is one of these features.

The Samples.
Fifteen samples (M1-M15) were collected (Figures 2-3) and were analyzed by ICP-MS (Table 1).Some samples (M1-M12) were taken from the button to the top in Pit 6 (Table 1 and Figures 2 and 3).Sample M1 was collected, 12 cms above the bad rock.This is apparently clayey silt with some organic content to judge by its color (dark gray).Sample M2 was taken 12 cms above M1.It is composed by darkcolored clayey silt with some gravel intrusions.Sample M3 was taken 10 cms above M2 and it is formed by mudstones intrusion with very dark color.Sample M4 was taken 25 cms above M3 and it is composed of a bundle of gray silt with gravel intrusions.Sample M5 was collected 15 cms above M4 and it is composed by a horizontal band above clayey silt with high organic content judging by its dark color.Sample M6 is located 10 cms above M5 and it is composed of a package of clear silt, with abundant intrusions of loam blocks, and appears to have been originated from a loosening of Pit 6 walls.The sample M7 was collected 45 cms above M6 and it is composed of a package of dark silt with somewhat of organic matter.Sample M8 is situated 70 cms above M7, a package of dark silts that may correspond to one of the paleosols documented filling this gap.Sample M9 was taken 40 cms above M8 and is composed of a package of silt with little intrusion.
Sample M10 (UE104355) was located 25 cms above M9.This sample is composed of a darkish silt package that seems to correspond to a deposition period.In this stratigraphic unit (UE104355) a large core of flint was found indicating that this sample was collected in a unit where anthropogenic  components are present.Sample M11 was collected 75 cms above sample M10 and is formed by a clear silt package that interposes a well-documented paleosol.Sample M12, 20 cms above the previous M11, is composed of dark silt package that is associated with a paleosol that fills an earlier erosion period dated after the ca.5000 cal.BC.Sample M13 is marl in contact with the UE104355 (M10).The M14 (marl) has been directly taken from the parent rock.The M15 sample was sampled in the sector BC105 and it is composed by a dark area and silty sediment from the Bronze Age dated by the recovered ceramic fragments.

Chemical Analysis of Soil.
The sampling was carried out collecting the sediments sequentially from Pit 6 stratigraphic wall (Figures 2-3) and not sieved.The 15 samples have been sampled using microspoon spatula made of stainless steel, always cleaned before taking a new sample, and stored in 15 mL test tubes.Sample preparation and digestion have been carried out as described by Gallello et al. [1].The process includes a thermal treatment at 450 ∘ C. Successively the samples were crushed and homogenized employing an agate mortar.A partial soil analysis has been carried out and we have developed a digestion method and a range of dilutions from the digested solution in order to provide reproducible and comparable results compatible with the sensitivity of the analytical method employed.
The digestion method consisted of the addition of 1.5 mL HCl and 1.5 mL HNO 3 to 0.5 g of sample in glass tubes placing them in a water bath at 100 ∘ C for 40 min.In addition, two blanks were prepared using the same acid volumes and digestion conditions of the samples.Subsequently, the digested solutions have been carefully poured into plastic tubes of 15 mL, bringing the volume to 15 mL with purified water.This concentrated solution (A) was diluted 1 : 10 obtaining a solution (B) to measure La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu, Sc, and Y. Dilution containing 10% concentrated HCl and 10% concentrated HNO 3 has been maintained in all solutions.A multielemental stock solution containing REE, Sc, and Y at concentration of 100 g/mL was prepared.The calibration standards were prepared employing 50 mL volumetric flasks and adding 5 mL HCl + 5 mL HNO 3 plus the corresponding volume of standard solution and filling up to volume with pure water.Solutions were analyzed by ICP-MS with Perkin Elmer Elan DRCII (Concord, Ontario, Canada).To avoid the obstruction of the nebulizer system the sample was filtered employing filter paper (Whatman N.1 of 70 mm).Concentrations ranging between 0 and 0.6 g/mL have been used for La, Ce, Pr, and Nd and concentrations ranging between 0 and 0.1 g/mL for Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, and Lu.All standards were acquired from Sharlab S.L. (Barcelona).The standard error of readings during the analysis ranged from 3% to 9% for REE.Soil GBW07408 has been used as standard reference material for evaluating the analytical method.Rh was used as internal standard.REE concentrations were normalized to Post-Archaean Australian Shales (PAAS) concentrations using the values of Taylor and McLennan [28].2. The results expressed in g/g are shown in Table 3.The samples (from M1 to M12) have been sequentially collected from the bottom to the top in Pit 6 (Figures 2 and  3).The sample M13 is marl situated in contact to UE104355 where the sample M10 was taken.The sample M14 is marl belonging to another excavation sector.Also the sample M15 belongs to an archaeological stratigraphy located in a different excavation sector dated in bronze age (BC105).Observing the results (Table 3) M1 contains the lowest values of La, Ce, Nd, Eu, Tm, Yb, Sc, and Y between the samples.Gradually REE values arise sequentially between M2, M3, and M4.Samples M5, M6, M7, M8, and M9 maintain lanthanides ).Finally the sequence sampled in Pit 6 reveals that the anthropogenic units (M10) and the directly related deposits (M13) show a REE enrichment.

Relation of REE/REE PAAS .
Ratios La/Yb PAAS , La/Gd PAAS [15], La/Sm PAAS , and Sm/Yb PAAS [29] were used to characterize the REE/PAAS profiles in terms of relative enrichment or depletion of light REE (La-Nd), medium REE (Sm-Ho), and heavy REE (Er-Lu).Table 4 shows differences in La/Yb PAAS , La/Gd PAAS , La/ Sm PAAS , and Sm/Yb PAAS ratio between the samples similar to those observed in Table 3 Summing up, M10 contains the highest values of La/Yb PAAS , La/Gd PAAS , La/Sm PAAS , and Sm/Yb PAAS and a mayor enrichment of the sequence (M1-M12).A moderate reduction in LREE, MREEs, and HREE can be observed for the intermediate samples of the sequence (M2-M3-M4-M5-M6-M7-M8-M9-M11).A mayor reduction in LREE, MREE, and HREE can be observed at the edge of the sequence, at the bottom (M1), and at the top (M12).Probably this   The " * " and the  subscript denote the anomaly and the concentration of standard Post-Archean Australian Shale (PAAS).
In this set of samples Eu * and Ce * values are less than one (Figure 4) and show a strong negative europium anomaly and cerium negative anomaly.That could point to more reducing conditions, without oxygen access in the archaeological site, proving stability in the studied stratigraphic units.The same Eu * and Ce * results have been previously obtained at the same site [1].

Conclusion
ICP-MS analysis has shown to be an effective technique to observe objective internal relation between samples to test our hypotheses.We consider that REE analyses are useful to identify anthropogenic patterns due to the particular chemical characteristics of those elements.Our data show that REE may be good discriminators when there are not only important, but also hidden differences between stratigraphic units.The results of samples sequentially taken in Pit 6 have revealed that the sample M10 (UE104355) contains the greatest enrichment in REE of the entire sequence (M1-M12) and confirm the anthropogenic origin of this unit.A moderate reduction in LREE, MREE, and HREE values can be observed in intermediate samples of the sequence (M2-M3-M4-M5-M6-M7-M8-M9-M11).A reduction of values on the edge of the sequence (bottom sample M1 and top sample M12) may reflect the natural origin of these units.Analysis of samples taken sequentially in a huge section (Pit 6) shows that rare earth elements may help to identify areas influenced by a higher human impact and to define the degree of anthropization on the environment also in a range of few centimetres.
Poor preservation of ancient remains in archaeological stratigraphy could carry uncertainty to identify anthropogenic traces.REE analysis can indicate whether the formation of soil is due to human activities or natural processes.Many techniques may be able to shed light on ancient human group activities, but we are confident that the commonly employed methodologies can be enhanced by the use of REE.Our approach allows us to evaluate the degree of human contribution to the formation of paleosols, providing new ways to deal with one of the most classic archaeological problems.
Certainly, more analyses of rare earths are needed to better characterize the composition of the parent rock and to gain more knowledge about the dynamics that take place to differentiate natural and anthropogenic soils, not only in the archaeological area of Mas d'Is but also at regional level.

Figure 1 :
Figure 1: Location of the upper and middle Serpis Valley.

3. 1 .
Composition of Samples.15 samples have been analyzed by ICP-MS to determine REE, Sc, and Y contents.The analyzed isotopes, instrumental LOD and LOQ, and  2 are listed in Table expressed in g/g.
Note: Instrumental LOD and LOQ.* Internal standard.are the latest belonging to the sequence and show low REE concentrations in relation to the rest of samples except M1 which maintains the lowest levels.Sample M14 (marl) belongs to an excavation area far from Pit 6. REE values in this sample (M14) are in general similar to the sequence samples which contain the lowest values.Sample M15 dated in Bronze Age was taken in another excavation sector (BC105); this sample contains REE values that are similar to the average of the sequence (M1-M12 . M1 contains the lowest values of La/Yb PAAS , La/Gd PAAS , La/Sm PAAS , and Sm/Yb PAAS .Those values arise gradually trough the samples M2, M3, and M4.Samples M5, M6, M7, M8, and M9 show values similar to M2.The sample M10 (UE104355), rich in archaeological materials, presents again the highest La/Yb PAAS , La/Gd PAAS , La/Sm PAAS , and Sm/Yb PAAS values.Sample 13 (marl) related to M10 shows values slightly lower than M10.Samples M11 and M12 are the latest belonging to the sequence and show lower REE concentrations in relation to the rest of samples except M1 which maintains the lowest levels.Sample M14 (marl) values are in general similar to the sequence samples which contain the lowest values.Sample M15 (Bronze Age) contains again values similar to the average of the sequence (M1-M12).

Table 3 :
Rare Earth Elements sample concentrations and their standard deviations (±st/√).