A geochemical multi-proxy approach for anthropogenic processes in aMiddleeUpper Pleistocene endokarstic deposit
Guadalupe Monge a, *, Francisco J. Jimenez-Espejo b, Manuel Pozo c, María I. Carretero a,Cecilio Barroso d
a Dpto. Cristalografía, Mineralogía y Q. A. Facultad de Química, Universidad de Sevilla, Spainb Department of Biogeochemistry, Japan Agency for Marine-Earth Science and Technology, Yokosuka, Japanc Dpto. Geología y Geoquímica, Universidad Aut�onoma de Madrid. Spaind Fundaci�on Instituto de Investigaci�on de Prehistoria y Evoluci�on Humana. Lucena. Spain
Deciphering human activities in archaeological sites is a priority issue in archaeological studies, never-theless its geochemical fingerprints on sediments are poorly known. In sites belonging to the recentprehistory these geochemical signals have been taken into account, but in oldest sites this subject has notbeen studied sufficiently.
The aim of this paper consists on tracking geochemical proxies that can be attributed to anthropogenicprocesses in endokarstic Pleistocene deposits. Recognize these elements can be a key factor in order toexplore the potential of non-excavated archeological levels and find out activities performed in thosesediments more accurately. For that purpose a MiddleeUpper Pleistocene endokarstic deposit (Cueva del�Angel) belonging to the Iberian Peninsula has been chosen. This site provides numerous evidences ofhuman activities, as butchering and cooking of predated animals or the habitual use of fire throughout itsmain stratigraphic sequence.
This geochemical/archaeological approach highlights that the upper units consist of anthropogenicinfluenced sediments, while the lower unit shows a greater percentage of geogenic inputs. Based on Pand ZneCueSr, several levels with higher anthropogenic inputs have been identified. These two attri-butes can be suggested as proxies of human activities for this site. High values of P appear to be linkedwith “butchering highly occupied” levels, and high levels of ZneCueSr seem to be related with fires. Thisgeochemical information has been compared and tested with previous archeological information.
Cave deposits are one of the most important archives in thegeological record to infer past events, including anthropogenicones, as they are unique environments preserving sedimentsderived from an assortment of geological and human processes(Goldberg and Sherwood, 2006). They can provide not only sub-stancial information on the climatic and geomorphological historyof the cave itself and its surroundings (Karkanas and Goldberg,2013), but also a wealth of contextual information for interpret-ing the archaeological remains and the human role in the formationof the endokarstic deposit (Goldberg and Sherwood, 2006).
reserved.
Although it has long been recognized that the study of artifactswithout regard to their context is of limited value in archaeologicalinterpretation (Schiffer, 1972, 1983), traditionally the attention hasfocused on archaeological remains like lithic tools or fossil bones,and only recently has systematic classification of cave sedimentsbeen proposed (Ford and Williams, 2007; White, 2007; Trappe,2010).
The clue for a correct overall palaeoanthropological interpreta-tion of a sitewill be a suitable characterization of cave sediments, aswell as being able to differentiate geogenic sediments fromanthropogenic ones. The geochemical characterization of sedi-ments could be so helpful on deciphering the “anthropogenic de-gree” of a sedimentary level and deposit evolution, beingcommonly applied in lacustrine environments (García-Alix et al.,2013) and historical archaeological sites (Kawahata et al., 2014).In prehistoric sites, macroscopic physical features still represent the
G. Monge et al. / Quaternary International 407 (2016) 140e149 141
primary evidence of human occupation, although their distributioncould be very patchy as many activities produce few (if any), whilegeochemical indicators which are directly related to occupationitself are more homogeneously deposited (Schlezinger and Howes,2000; Monge et al., 2015) (e.g. ashes).
The overall relation between archaeology and geochemistry canbe described as the enrichment or depletion of certain elements insediments through the act of human occupation (Oonk et al., 2009).Traditionally, geochemical analysis of sediments have been used toconfirm, deny, or expand the results achieved through other tech-niques (Popenoe, 1959; Cowgill, 1961; Cowgill and Hutchinson,1963). Recently, the rise of high resolution non-destructive tech-niques (e.g. XRF-Scanners) combined with equipment portability(e.g. portable LIBS) and price reduction, offered the possibility toapplicate these methods to climate and archaeological records (e.g.Marwick, 2005; Finlayson et al., 2006;Wilson et al., 2008; Skaberneet al., 2015). Geochemistry has become not only a characterizationtool, but also an exploration one, as a possible primary step in thedevelopment of an excavation strategy (e.g. Parnell et al., 2001;Wilson et al., 2008).
These advances and the difficulty of the interpretation ofgeochemical data as proxies (Wilson et al., 2008), justify the re-evaluation of the relationship between geochemical and archaeo-logical data, and between different elements as indicators.
The relatively low number of sites that have used geochemistryto address archaeological questions, include the Amerindian siteCape Cod (Schlezinger and Howes, 2000), where a comprehensivestudy of an anthrosol from the last glacial period reveals the utilityof organic P and elemental ratios in delineating human occupationsin sandy, acidic soils. Also, the prehistoric archaeological sites atBan Non Wat and Nong Hua Raet in Thailand (Kanthilatha et al.,2014), from at least 4000 BP, revealed P, Ca an K as key anthropo-genic elements which reflect the occupation intensity of ancientpeople in different floor surfaces.
In Guatemala, during Maya period, two sites has beengeochemically studied. Piedras Negras site (Parnell et al., 2002),where elevated levels of Ba, P and Mn were found to be associatedwith areas of organic waste disposal whilst Hg and Pb concentra-tions were associated with craft production areas. At Las Pozas(Fern�andez et al., 2002), high levels of P, K, Mg and pH were relatedwith food preparation areas, as well as high P concentrations and
Table 1Mainly descriptive and mineralogical features from J/K stratigraphic levels and location
low pH with food consumption areas. During historical periods,geochemical studies mainly focused on Pb, Hg and other metal-related smelting activities (Kawahata et al., 2014), pigments(Emslie et al., 2015) or farming activities that can be recognized byCa, Sr, P, Zn and Cu concentration patterns linked to charcoal andbone mediated for late 1800s farms (Wilson et al., 2008).
As geochemistry has not been used systematically at old pre-historic contexts (Oonk et al., 2009), this study focus on decipher-ing geochemical proxies which could be used to identify humanactivities. For that purpose Cueva del �Angel (Spain) has been chosenbecause it provides numerous macroscopic evidences of humanactivities (mainly fossil bones and lithic tools along the strati-graphic sequence). This paper also aims to review different ele-ments and consider their validation as anthropogenic/geogenicproxies related to archeological sites located in caves.
2. Cueva del �Angel
Cueva del �Angel is located in the south of the Iberian Peninsula,near the town of Lucena in the province of C�ordoba (Spain). Thecave is situated at an altitude of 620 m above sea level (37�220N,4�280W) on the foothills of the Sierra de Araceli (Fig. 1).
From a geological point of view, this cavity is hosted in aMesozoic carbonate unit composed of limestone and dolostones(Lower and Middle Lias), belonging to the Betic Ranges (L�opez-Chicano, 1990). Nowadays, the roof and walls are partiallycollapsed. Thus, the archaeological site is located on an open-airplatform measuring around 300 m2 with a strong slope south-wards (Monge et al., 2014).
The most complete stratigraphic profile, named J/K, presentstwenty stratigraphic levels organized in three stratigraphic units.They have been differentiated on the basis of lithology, colour,texture, structure, coarse fraction percentage, porosity and occur-rences of archaeological material. Detailed stratigraphic de-scriptions and mineralogical results can be found in Botella et al.(2006), Barroso et al. (2011), Monge (2012) and Monge et al.(2014). The major traits are summarized in Table 1.
In a previous study, Barroso et al. (2011) state that the faunalassemblage dominated by Equus ferus, large bovids and cervids hasbeen subjected to intense human actions reflecting selective pre-dation (fragmentation of the bones for marrow extraction with an
of geochemical samples at sedimentary levels.
G. Monge et al. / Quaternary International 407 (2016) 140e149142
appreciable number of cutmarks and striations related to deflesh-ing, filleting and disarticulation). Besides, the percentage of burntbones in the sequence is testimony of the intensive use of fire. Butat present, no hypothesis on the various modes of use of fire ispossible. Preliminary dating suggests a long sequence between MIS11 and MIS 5.
3. Material and methods
Taking into account the distribution, arrangement and thicknessof the stratigraphic levels within profile J/K, a total of 48 sampleswere selected for their geochemical characterization (Table 1).
The geochemical analysis of the samples was carried out atActivation Laboratories Ltd., Ontario (Canada) for a total of 64 ele-ments analysed by several techniques (INNA:Instrumental Neutron
Table 2Main geochemical elements in sediment samples and ashes from experimental fires.
Activation Analysis; TD: Total Digestion; ICP: Inductively CoupledPlasma and MS: Mass Spectrometry).
In order to analyze the possible relationships between thedifferent geochemical variables, a statistical approach using thestatistical package Statistica Statsoft 9 was conducted. After elimi-nation of those variables with no information (values belowdetection limit) or repeated ones, scatter and linear regression di-agrams, Pearson coefficient (R), principal component analysis andcluster analysis according to Ward's method (Ward, 1963) werecarried out. For thirty-three of the profile samples (marked withasterisk in Table 2), carbonate d13C and d18O were analysed for theirstable isotopes as well as seven samples from the host rock andspeleothems (Fig. 2).
For the determination of stable isotopes an Isotope Ratio MassSpectrometer THERMO DELTA V Advantage has been used. Theanalyses were performed on an aliquot of sieved samples at
Fig. 2. Situation of the speleothems (I) and host rock (HR) samples in the archaeo-logical site.
Fig. 1. Map showing the geographical location of Cueva del �Angel, Lucena (C�ordoba).
G. Monge et al. / Quaternary International 407 (2016) 140e149 143
Laboratorio de Is�otopos Estables del Servicio Interdepartamental deInvestigaci�on (SIDI) of the Universidad Aut�onoma de Madrid(UAM). The carbonates were analyzed by reaction with H3PO4 103%and 25 �C, in a line of high vacuum, following the classical princi-ples of McCrea (1950), and applying the correction of Craig (1957).
The results have been expressed in‰ delta notation, relative to thestandard VPDB (Vienna Pee Dee Belemnite).
The isotopic signals of the sediment samples have beenanalyzed in order to discriminate the contribution of geogeniccalcite with respect to anthropogenic one, according to the meth-odology proposed by Shahack-Gross et al. (2008).
As previous studies (Wilson et al., 2008;Monge et al., 2015) haveidentified Zn, Cu and Sr as enriched elements in ashes from fireswhen compared with other background soil geochemical infor-mation, a ZneCueSr Factor has been calculated from ash data fromexperimental fires (Monge et al., 2015), based on enrichment factorestimations (Adamu and Nganje, 2010). The used divisors for eachelement were calculated from average concentrations of theexperimental data given by Monge et al. (2015) as follows (fordetails see Table 2):
Zn� Cu� Sr Factor ¼ ½ðZn=495Þ þ ðCu=115Þ þ ðSr=390Þ�
where:
- Zn, Cu and Sr are the contents of these elements of each sedi-ment sample,
G. Monge et al. / Quaternary International 407 (2016) 140e149144
4. Results
From the statistical point of view, the analysed geochemicalelements yield four different associations (Fig. 3) with strong directdependences within each association (Pearson coefficients > 0.7):
� Association 1 (AleFeeKeYeCseRbeTheSceLaeSmeEueLu):There is a tendency towards a gradual increase in those ele-ments concentrations (Fig. 3) from depthe 450 cm (level XV), toachieve the maximum values at e550 cm (level XX).
� Association 2 (HfeZreNbeTi): These elements show low con-centrations values, with occasional increases in which thevalues reached triple or even quadruple the averagevalues (Fig. 3). These increases are reflected at depths: �280 cm(level III),�355 cm (level VI),�421 cm (level X), and levels XVIII,XIX and XX.
� Association 3 (CueNi): This group presents greater homogene-ity than previous associations. Almost all the values aredistributed within a defined range between 10 and 45 ppm,
Fig. 3. Plot showing different associations of geochemical elements concentrations vsdepth.
except for some specific levels with a significant increase (UnitII) (Fig. 3).
� Association 4 (PeNaeSr): High values of these elements arerecorded both in stratigraphic unit I (with punctual increases atlevels III and VI), and II (with punctual increases at levels VIII, IX,XIII, XIV and XV). Otherwise, their values decrease sharply atstratigraphic unit III.
The relations between all the samples were examined by meansof cluster analysis. The samples were grouped into clusters byWard's method (Fig. 4). This analysis reports the existence of twogroups of samples related with the values of the variables.
With regard to the values of the stable isotopes analysed d13Cand d18O (Fig. 5), sediment samples show mean values of �10.08‰d13C (ranging between �8.96‰ and �10.77‰ d13 C) and �5.2‰d18O (ranging between �4.7‰ and �5.7‰ d18O). d13C values seemto present a moderate homogeneity, except from V and XV levels,while d18O values show a tendency of increase with depth, mainlyin stratigraphic unit III. It ought to be highlighted the increase ofd13 C value in the transition zone between Units II and III. Theisotopic signal of speleothems, with d 13C values rangingbetween �8.90‰ and �10.88‰ and �5.8‰ and �7.2‰ for d18O arenormal within the range of values accepted for these lithologies(Sharp, 2007).
Applying the methodology proposed by Shahack-Gross et al.(2008) in order to discriminate the contribution of geogeniccalcite with respect to anthropogenic, the isotopic signal betweentwo extreme members have been compared: in our case the iso-topic signal of the speleothems and host rock (which is taken as theisotopic signal of the sediment if all the calcite in it was geogenic),with the isotopic signal of modern ashes from bushes C-3: oak(d13C ¼ �22,37‰, d18O ¼ �16,44‰) and carob (d13C ¼ �24,54‰,d18O ¼ �17,33‰).
It is noted (Fig. 6) that samples from the stratigraphic profile andspeleothems have a different isotope signal suggesting that calcitepresent in sediment samples has not only a geogenic origin.Nevertheless, the absence of intermediate terms between the iso-topic signal of the speleothems and C-3 vegetation indicate a minorinfluence by the vegetation input. This could be linked with aninfluence by bone dissolution (Karkanas et al., 2007). In addition, ithas not been possible to infer any relationship between the isotopicvalues of the sediment samples and the different stratigraphicunits.
Fig. 4. Cluster analysis from sediment samples of J/K profile (Ward's method).
Fig. 5. d13C and d18O values from sediment samples along J/K stratigraphic profile (a), speleothems and host rock samples (b).
Fig. 6. d 13C and d18O values in speleothems, host rock, sediment samples and vegetation samples.
G. Monge et al. / Quaternary International 407 (2016) 140e149 145
5. Discussion
Elements belonging to Association 1(AleFeeKeYeCseRbeTheSceLaeSmeEueLu) reflect a moderatepositive correlation (R ¼ 0.6808) with phyllosilicates (Table 1).
Previous mineralogical studies at the site (Monge, 2012) demon-strate that the provenance of phyllosilicates come from externalcontributions or are related with the insoluble residue of limestoneand dolomite, confirming the mainly inherited character of theelements belonging to this association.
G. Monge et al. / Quaternary International 407 (2016) 140e149146
Elements from Association 2 (HfeZreNbeTi) have been relatedwith the presence of heavy minerals in different environments (e.g.Rodrigo-G�amiz et al., 2011; Bahr et al., 2014; Jim�enez-Espejo et al.,2014). Heavy minerals standing out in J/K profile (Monge, 2012)include grains of disthene, rutile, andalusite, zircon and biotite, butalso garnet, titanite, tourmaline and hornblende can be found. AtCueva del Angel site, they can be related to the insoluble residue oflimestone and dolomite or with aeolian contributions coming fromthe Triassic “olistostrome” unit surrounding the cave. Sedimentary,volcanic, and metamorphic minerals has been described in thism�elange unit (Rold�an et al., 2012).
Association 3 (CueNi) are redox sensitive elements (Tribovillardet al., 2006) and their mobility in soils is related to organic mattercontent, element speciation, water content among others (Fic andSchroter, 1989; Kiikkil€a et al., 2002). In general, heavy metals areconsidered relatively immobile in soils (Ross, 1996). In some typesof soils the presence of carbonates effectively immobilizes Cu byproviding and adsorbing surface or by buffering pH (Dudley et al.,1991). Nevertheless their mobility can be enhanced when theseheavy metals form complexes with soluble organic matter in thesoil percolation water (Li and Shuman, 1997). As Cueva del Angelsite has not suffer extensive diagenesis processes (Monge, 2012),these elements seem to be related to the use of fire (discussedbelow) (Monge et al., 2015).
Focusing on Association 4 (PeNaeSr), anthropogenic processesare inferred because P (phosphorous) is the most widely usedanthropogenic indicator in archaeological sites (Oonk et al., 2009).Such enrichment can be also characteristic of guano deposits orintense carnivore actions (Skaberne et al., 2015), but in this casethose factors can be discarded (Barroso et al., 2011). Indeed, asaccumulations of guano are possible only during non-occupationalperiods (Courty et al., 1989) and intense carnivore signals have notbeen found, this site is supposed to be a continuous habitat forHomo during a long period of time. As the other two elementsdisplay a strong direct correlation (Pearson coefficients: RP-
Na ¼ :0.78; RP-Sr ¼ 0.76) with P, they are supposed to be also relatedwith human activities.
Despite the important number of variables used (major ele-ments, minor elements, trace elements and rare earth elements-REE) in order to tracking down possible proxies of human occu-pation on the site, the statistical analysis yielded two associationswhich can be related with human actions. As previous studies havedemonstrated anomalous concentrations of P, Ca, Mg, Cu, Zn and insome cases Sr, as the most commonly elements found in archaeo-logical soils (Cook and Heizer, 1965; Eidt, 1984; Ottaway andMatthews, 1988; Middleton and Price, 1996; Haslam and Tibbett,2004), we emphasize them as possible geochemical proxies forhuman activities in this endokarstic deposit.
The interpretation of element concentration patterns inarchaeological sites is problematic due to the effects of post-depositional processes, which may influence the retention andredistribution of anthropogenic elements hosted in the sedi-ments (Wilson et al., 2008). However, the relatively confined andstable environmental conditions in endokarstic deposits(without subaerial exposition to atmospheric weathering pro-cesses) helps to minimize these processes if we compare themwith open air sites (e.g. Ji et al., 2004). At the same time, thegeochemical fingerprints of this type of environments reflecthigh pHs and continuous contributions of carbonated waterscoming from limestone and dolostone dissolution. These twofactors: high pHs and intense carbonation of the sediments willcontribute to fix relatively mobile elements along the profile. Foreach element, the validation as an anthropogenic/geogenicproxy related to archaeological sites within caves is discussedbelow:
5.1. Phosphorous (P)
P is the most widely used anthropogenic indicator in archaeo-logical sites (Oonk et al., 2009). P is normally prevalent in plant andanimal tissue, bones, urine, faeces and ashes, and thus it is a keyelement in occupation waste and manure (Proudfoot, 1976; Bethelland Mate, 1989). In addition, P is relatively stable in most soils(Robinson et al., 1995; Tiessen, 1995) and sediments. The generalconsensus is that most anthropogenic P in soils becomes inmobi-lized shortly after deposition (Robinson et al., 1995; Gale et al.,2000; Hooda et al., 2001; Oonk et al., 2009). Likewise relatedwith post-depositional processes (carbonation of the profile), cal-cium serves to fix phosphate compounds for long-time periods(Parnell et al., 2001).
Statistical analysis of P contents and macroscopic character-istics of human activities as the presence of fossil bones (B) orlithic tools (LT) along the J/K profile (Fig. 7 and Table 1) reflectstrong dependences (RP-B ¼ 0.77; RP-LT ¼ 0.81). The higher lithictools and fossil bones densities were found at depths where Pwas enriched, validating P as an important proxy of human ac-tivities in the site.
When crossing data from Association 4 (PeNaeSr) with Clusteranalysis, the samples with a higher P content (>3.95%) are thoseforming one of the groups from the Cluster analysis (Fig. 4). Thus,the statistical study support that P content is related with sedi-mentary beds showing the highest levels of occupation intensity(Schlezinger and Howes, 2000; Wells et al., 2000; Parnell et al.,2002; Marwick, 2005). Samples with P values > 3.95 due to Clus-ter analysis are located at levels III, VI, VIII, IX, XIII, XIV and XV(Fig. 7).
5.2. Calcium (Ca) and Magnesium (Mg)
Although Ca have been linked in different studies (Wilson et al.,2008; Kanthilatha et al., 2014), as indicator of human activities inarchaeological sites, at Cueva del �Angel site, most of the Ca of thesamples is related to calcite content (R ¼ 0.6851). As previousmineralogical studies at Cueva del �Angel have demonstrated(Monge, 2012), calcite origin results from calcium carbonate pre-cipitations within the cave, due to karsitfication of Liassic lime-stones and the formation of speleothems (authigenic calcite), butalso due to rock-falls from the roof and walls of the cave (in bothcases with a geogenic origin). Nevertheless it could have also ananthropogenic origin by carbonation of ashes generated during thecombustion of wood in hearths (Kennedy et al., 1968; Scurfieldet al., 1973; Humphreys et al., 1987; Schiegl et al., 1996; Karkanaset al., 2000; Weiner et al., 2002).
If the geochemical signal from geogenic calcite could be differ-entiate from the geochemical signal of anthropogenic calcite, itcould be used as a geochemical proxy of human activities, but as itwas exposed previously, it can be concluded that in this kind ofenvironment (endokarstic deposits), Ca is an unsuitablegeochemical proxy of human activities because it is ubiquitous insuch environments.
Mg is not a good anthropogenic indicator in endokarstic de-posits where dolomite is part of the host rock (as is the case atCueva del Angel site), because it is not possible to distinguish itsanthropogenic geochemical signal from its geogenic one. The lackof evidence explains its absence in Association 4.
5.3. Copper (Cu), zinc (Zn) and Strontium (Sr)
Oonk et al. (2009) reported Cu as one of the most promisinganthropogenic indicators, because it is relatively stable in soils and
Fig. 7. Synthetic stratigraphic section from J/K profile with concentration vs depth of Fossil Bones, Lithic Tools, P and ZneCueSr Factor.
G. Monge et al. / Quaternary International 407 (2016) 140e149 147
sediments (Fontes and Gomes, 2003), while Zn stabilization mainlyoccurs upon interactions with phyllosilicates (Cao et al., 2004).
Zn and Cu sediment content can be used as anthropogenicproxies related with hearths when guano inputs (Miko et al., 2002)and/or extensive diagenesis can be discarded (Monge et al., 2015).Patterns of these elements along the entire levels where hearthswere found can be attributed to fly ash and/or wood ash redistri-bution. As Cueva del Angel does not reflect extensive diagenesis(Monge, 2012), or guano deposits (Barroso et al., 2011), Zn and Cu(Association 3) could be related to the evidence of the use of firereported at the site.
If we compare the macroscopic characteristics of human activ-ities as the presence of Fossil Bones (B) or Lithic Tools (LT) along J/Kprofile with Cu, Zn or Sr separately, no relationships are noteworthy(only for Sr- Lithic Tools RSr-LT ¼ 0.74). If we calculate the ZneCueSrFactor and compare itwith Lithic Tools (RCu-Zn-Sr-LT¼ 0.67) and FossilBones (B) (RCu-Zn-Sr-B ¼ 0.79) the data displayed elevated de-pendences (Fig. 7), the highest one related with the presence ofbones. The huge percentage of burnt bones in the sequence suggeststhat by combining Cu, Zn and Sr in a single factor (ZneCueSr factor),the data yield a robust tracer for calculating the anthropogenicdegree of each sedimentary level. Related to the use offire at the site,fossil bones show higher correlations with the obtained geochem-ical proxies because lithic tools distribution in sedimentary se-quences is normally patchy (Schlezinger and Howes, 2000). Indeed,it seems that fossil bones have been used as fuel at the site (Barrosoet al., 2011), and comparing P content with the ZneCueSr Factor(RCu-Zn-Sr-P ¼ 0.80) a strong direct dependence was found (Fig. 7).
6. Conclusions
Although more studies are needed to reveal the importance ofanthropogenic geochemical proxies in endokarstic deposits, this
study has shown clearly that specific features have the potential forthe application as geochemical proxies and the use as indicators ofhuman activities. The most reliable geochemical proxy is P, whilethe calculated ZneCueSr Factor is a robust tracer related to hearthremains. Ca and Mg are discarded for this type of environment,because their possible anthropogenic fingerprint strongly mergeswith their geogenic signal.
Based on P content and ZneCueSr Factor, at Cueva del �Angelstratigraphic units I and II show a higher anthropogenic contribu-tion (specifically at levels III, VI, VIII, IX, XIII, XIV and XV). The basalstratigraphic unit III presents more homogeneous values related togeogenic inputs. The archaeological information provided by thesite corroborates the geochemical multi-proxy approach.
Acknowledgements
This work was funded by the “Instituto de Investigaci�on de Pre-historia y Evoluci�on Humana” Foundation, UAM C-144 ResearchGroup and the Andalusian Research Group Board RNM-349.
References
Adamu, C.I., Nganje, T.N., 2010. Heavy metal contamination of surface soil in rela-tionship to land use patterns: a case study of Benue state, Nigeria. MaterialsSciences and Applications 1, 127e134.
Bahr, A., Jim�enez-Espejo, F.J., Kolasinac, N., Grunert, P., Hernand�ez-Molina, F.J.,R€ohl, U., Voelker, A., Escutia, C., Stow, D., Alvarez-Zarikian, C.A., Hodell, D.A.,2014. Deciphering bottom current strength and paleoclimate signals fromcontourite deposits in the Gulf of C�adiz during the last 140 kyr: an inorganicgeochemical approach. Geochemistry, Geophysics, Geosystems 15 (8),3145e3160.
Barroso, C., Botella, D., Caparr�os, M., Moigne, A.M., Celiberti, V., Testu, A., Barsky, D.,Notter, O., Riquelme, J.A., Pozo, M., Carretero, M.I., Monge, G., Khatib, S., Saos, T.,Gregoire, S., Bail�on, S., García-Solano, J.A., Cabral, A.L., Djerrab, A., Hedley, I.G.,Abdessadok, S., Batalla, G., Astier, N., Bertin, L., Boulbes, N., Cauche, D., Filoux, A.,Hanquet, C., Milizia, C., Moutoussamy, J., Rossoni, E., Verdú, L., de Lumley, H.,
G. Monge et al. / Quaternary International 407 (2016) 140e149148
2011. The Cueva del Angel (Lucena, Spain): an Acheulean hunters habitat in theSouth of the Iberian Peninsula. Quaternary International 243, 105e126.
Bethell, P., Mate, I., 1989. The use of soil phosphate analysis in archaeology: acritique. In: Henderson, J. (Ed.), Scientific Analysis in Archaeology and itsInterpretation. U.C.L.A. Institute of Archaeology, Los Angeles, pp. 1e29.
Botella, D., Barroso, C., Riquelme, J.A., Abdessadok, S., Caparr�os, M., Verdú, L.,Monge, G., Solano, J.A., 2006. La Cueva del �Angel (Lucena, C�ordoba), un yaci-miento del Pleistoceno medio y superior del sur de la Península Ib�erica. Tra-bajos de Prehistoria 63, 153e165.
Cook, S.F., Heizer, R.F., 1965. Studies on the Chemical Analysis of ArchaeologicalSoils. University of California Press, Los Angeles.
Courty, M.A., Goldberg, P., Macphail, R., 1989. Soils and Micromorphology inArchaelogy. Cambridge University Press, Cambridge.
Cowgill, U.M., 1961. Soil fertility and the ancient Maya. Transactions of the Con-necticut Academy of Arts and Sciences 42, 1e56.
Cowgill, U.M., Hutchinson, G.E., 1963. Ecology and geochemistry of the MayaLowlands. Southwestern. Journal of Anthropology 19, 267e286.
Craig, H., 1957. Isotopic standards for carbon and oxygen and correction factors formass spectrometric analysis of carbon dioxide. Geochimica et CosmochimicaActa 12, 133e149.
Dudley, L.M., McLean, J.E., Furst, T.H., Jurinak, J.J., 1991. Sorption of cadmium andcopper from an acid mine waste extract by two calcareous soils: column study.Soil Science 151 (2), 121e135.
Eidt, R.C., 1984. Advances in Abandoned Settlement Analysis: Application to Pre-historic Anthrosols in Colombia. University of Wisconsin, Center for Latin:American Studies, Milwaukee.
Emslie, S.D., Brasso, R., Patterson, W.P., Carlos Valera, A., Mckenzie, A., Silva, A.M.,Gleason, J.D., Blum, J.D., 2015. Chronic mercury exposure in Late Neolithic/Chalcolithic populations in Portugal from the cultural use of cinnabar. ScientificReports 5, 14679. http://dx.doi.org/10.1038/srep14679.
Fern�andez, F.G., Terry, R.E., Inomata, T., Eberl, M., 2002. An ethnoarchaeologicalstudy of chemical residues in the floors and soils of Q'eqchi' Maya houses at LasPozas, Guatemala. Geoarchaeology 17, 487e519.
Fic, M., Schroter, M.I., 1989. Batch studies for the investigation of the mobility of theheavy metals Cd, Cr, Cu and Zn. Journal of contaminated hydrology 4, 69e78.
Finlayson, C., Giles, F., Rodríguez-Vidal, J., Fa, D.A., Gutierrez-L�opez, J.M., Santiago, A.,Finlayson, G., Allu�e, E., Baena-Preysler, J., C�aceres, I., Carri�on, J.S., Fern�andez-Jalvo, Y., Gleed, C.P., Jimenez-Espejo, F., L�opez, P., L�opez-S�aez, J.A., Riquelme, J.A.,S�anchez-Marco, A., Giles Guzman, F., Brown, K., Fuentes, N., Valarino, C.A.,Villalpando, A., Stringer, C.B., Martinez-Ruiz, F., Sakamoto, T., 2006. Late survivalof Neanderthals at the southernmost extreme of Europe. Nature 443, 850e853.
Fontes, M.P.F., Gomes, P.C., 2003. Simultaneous competitive adsorption of heavymetals by the matrix of tropical soils. Applied Geochemistry 18, 795e804.
Ford, D., Williams, P., 2007. Karst Hydrogeology and Geomorphology. Wiley andsons, Chichester.
Gale, P.M., Mullen, M.D., Cieslik, C., Tyler, D.D., Duck, B.N., Kirchner, M., McClure, J.,2000. Phosphorus distribution and availability in response to dairy manureapplications. Communications in Soil Science and Plant Analysis 31, 553e565.
García-Alix, G., Jimenez-Espejo, F.J., Lozano, J.A., Jim�enez-Moreno, G., Martinez-Ruiz, F., García Sanju�an, L., Aranda Jim�enez, G., García Alfonso, E., Ruiz-Puertas, G., Scott Anderson, R., 2013. Anthropogenic impact and lead pollutionthroughout the Holocene in Southern Iberia. Science of the Total Environment449, 451e460.
Goldberg, P., Sherwood, S.C., 2006. Deciphering human prehistory through thegeoarchaeological study of cave sediments. Evolutionary Anthropology 15,20e36.
Haslam, R., Tibbett, M., 2004. Sampling and analyzing metals in soils for archaeo-logical prospection: a critique. Geoarchaeology 19, 731e751.
Hooda, P.S., Truesdale, A.C., Edwards, A.C., Withers, P.J.A., Aitken, M.N., Miller, A.,Rendell, A.R., 2001. Manuring and fertilization effects on phosphorus accumu-lation in soils and potential environmental implications. Advances in Environ-mental Research 5, 13e21.
Humphreys, G.S., Hunt, P.A., Buchanan, R., 1987. Wood ash stone near Sydney, NSW:a carbonate pedological feature in an acidic soil. Australian Journal of SoilResearch 25, 115e124.
Ji, H., Wang, S., Ouyang, Z., Zhang, S., Sun, C., Liu, X., Zhou, D., 2004. Geochemistry ofred residua underlying dolomites in karst terrains of Yunnan-Guizhou Plateau.I. The formation of the Pingba profile. Chemical Geology 203, 1e27.
Jim�enez-Espejo, F.J., Garcia-Alíx, A., Jim�enez-Moreno, G., Rodrigo-G�amiz, M.,Anderson, R.S., Rodríguez-Tovar, F.J., Martínez-Ruiz, F., Giralt, S., Delgado-Huertas, A., Pardo-Igúzquiza, E., 2014. Saharan aeolian input and effective hu-midity variations over western Europe during the Holocene from a high altituderecord. Chemical Geology 374e375, 1e12.
Kanthilatha, N., Boyd, W., Chang, N., 2014. Multi-element characterization ofarchaeological floors at the prehistoric archaeological sites at Ban Non Wat andNong Hua Raet in Northeast Thailand. Quaternary International. http://dx.doi.org/10.1016/j.quaint.2014.07.067.
Karkanas, P., Bar-Yosef, O., Goldberg, P., Weiner, S., 2000. Diagenesis in PrehistoricCaves: the use of minerals that form in situ to assess the completeness of thearchaeological record. Journal of Archaeological Science 27, 915e929.
Karkanas, P., Shahack-Gross, R., Ayalon, A., Bar-Matthews, M., Barkai, R., Frumkin, A.,Gopher, A., Stiner, M., 2007. Evidence for habitual use of fire at the end of the
Karkanas, P., Goldberg, P., 2013. Micromorphology of cave sediments. In: Shroder, J.,Frumkin, A. (Eds.), Treatise on Geomorphology. Academic Press, San Diego, CA,pp. 286e297.
Kawahata, H., Yamashita, S., Yamaoka, K., Okai, T., Shimoda, G., Imai, N., 2014. Heavymetal pollution in Ancient Nara, Japan, during the eighth century. Progress inEarth and Planetay Science 1, 15. http://dx.doi.org/10.1186/2197-4284-1-15.
Kennedy, R.W., Satry, C.B., Barton, G.M., Ellis, E.L., 1968. Crystals in the wood of thegenus Abies indigenous to Canada and the United States. Canadian Journal ofBotany 46, 1221e1228.
Kiikkil€a, O., Derome, J., Brügger, T., Uhlig, C., Fritze, H., 2002. Copper mobility andtoxicity of soil percolation water to bacteria in a metal polluted forest soil. Plantand Soil 238, 273e280.
Li, Z.B., Shuman, L.M., 1997. Mobility of Zn, Cd nd Pb in soils as affected by poultrylitter extractdI. Leaching in soil columns. Environmental Pollution 95,219e226.
L�opez-Chicano, M., 1990. Sierra de Araceli y C.S.A. 1 (Lucena). Cuaderno de campo.In: AEQUA (Ed.), Vª Reuni�on de campo de AEQUA-Andalucía. Ayuntamiento deLucena, C�ordoba, pp. 56e64.
Marwick, B., 2005. Element concentrations and magnetic susceptibility of anthro-sols: indicators of prehistoric human occupation in the inland Pilbara, WesternAustralia. Journal of Archaeological Science 32, 1357e1368.
McCrea, J.M., 1950. On the isotopic chemistry of carbonates and a paleotemperaturescale. Journal of Chemical Physics 18, 849e857.
Middleton, W.D., Price, T.D., 1996. Identification of activity areas by multi-elementcharacterization of sediments from modern and archaeological house floorsusing inductively coupled plasma atomic emission spectroscopy. Journal ofArchaeological Science 23, 637e687.
Miko, S., Kuhta, M., Kapelj, S., 2002. Environmental baseline geochemistry of sed-iments and percolating waters in the Modri�c Cave, Croatia. Acta Carsologica 31,135e149.
Monge, G., 2012. Caracterizaci�on mineral�ogica, geoquímica y textural del rellenoPleistoceno de la Cueva del �Angel en Lucena (C�ordoba). University of Sevilla,Spain (Ph.D. thesis).
Monge, G., Carretero, M.I., Pozo, M., Barroso, C., 2014. Mineralogical changes in fossilbone from Cueva del �Angel, Spain: archaeological implications and ocurrence ofwhitlockite. Journal of Archaeological Science 46, 6e15.
Monge, G., Jimenez-Espejo, F.J., García-Alix, A., Martínez-Ruiz, F., Mattielli, N.,Finlayson, C., Ohkouchi, N., Cort�es, M., Bermúdez de Castro, J.M., Blasco, R.,Rosell, J., Carri�on, J., Rodríguez-Vidal, J., Finlayson, G., 2015. Earliest evidence ofpollution by heavy metals in archaeological sites. Scientific Reports 5, 14252.http://dx.doi.org/10.1038/srep14252.
Ottaway, J.H., Matthews, M.R., 1988. Trace element analysis of soil samples from astratified archaeological site. Environmental Geochemistry and Health 10,105e112.
Oonk, S., Slomp, C.P., Huisman, D.J., 2009. Geochemistry as an aid in archaeologicalprospection and site interpretation: current issues and research directions.Archaeological Prospection 16, 35e51.
Parnell, J.J., Terry, R.E., Golden, C., 2001. Using in-field phosphate testing torapidly identify middens at Piedras Negras, Guatemala. Geoarchaeology 16,855e873.
Parnell, J.J., Terry, R.E., Nelson, Z., 2002. Soil chemical analysis applied as an inter-pretive tool for ancient human activities in Piedras Negras, Guatemala. Journalof Archaeological Science 29, 379e404.
Popenoe, H., 1959. The influence of the shifting cultivation cycle on soil propertiesin Central America. In: Proceedings, Ninth Pacific Science Congress 7. Bangkok,pp. 72e77.
Proudfoot, B., 1976. The analysis and interpretation of soil phosphorus in archae-ological contexts. In: Davidson, D.A., Shackley, M.L. (Eds.), Geoarchaeology.Westview Press, Boulder, CO, pp. 93e113.
Robinson, J.S., Sharpley, A.N., Smith, S.J., 1995. The effect of animal manure appli-cations on the forms of soil phosphorus. In: Steele, K.F. (Ed.), Animal Waste andthe Lande Water Interface. CRC Press, Lewis Publishers, Boca Raton, FL,pp. 43e67.
Rodrigo-G�amiz, M., Martínez-Ruiz, F., Jim�enez-Espejo, F.J., Gallego-Torres, D., Nieto-Moreno, V., Romero, O., Ariztegui, D., 2011. Impact of climate variability in thewestern Mediterranean during the last 20,000 years: oceanic and atmosphericresponses. Quaternary Science Reviews 30, 2018e2034.
Rold�an, F.J., Rodríguez-Fern�andez, J., Villalobos, M., Lastra, J., Díaz-Pinto, G., P�erezRodríguez, A.B., 2012. Zonas: Subb�etico, Cuenca del Guadalquivir, Campo deGibraltar. In: Navas, J. (Ed.), GEODE. SIGECO. IGME. http://cuarzo.igme.es/sigeco.default.htm.
Ross, S.M., 1996. Retention, transformation and mobility of toxic metals in soils. In:Ross, S.M. (Ed.), Toxic Metals in Soil-plant Systems. John Wiley & Sons Ltd,Baffins Lane, Chichesten, pp. 63e151.
Schiegl, S., Goldberg, P., Bar-Yosef, O., Weiner, S., 1996. Ash deposits in Hayomin andKebara caves, Israel: macroscopic, microscopic and mineralogical observationsand their archaeological implications. Journal of Archaeological Science 23,763e781.
Schiffer, M.B., 1972. Archaeological context and systemic context. American Antiq-uity 37, 156e165.
Schiffer, M.B., 1983. Toward the identiication of formation processes. AmericanAntiquity 48, 675e706.
G. Monge et al. / Quaternary International 407 (2016) 140e149 149
Schlezinger, D.R., Howes, B.L., 2000. Organic phosphorus and elemental ratios asindicators of prehistoric human occupation. Journal of Archaeological Science27, 479e492.
Scurfield, G., Anderson, C.A., Silva, S.R., 1973. Silica in woody stems. AustralianJournal of Botany 22, 211e229.
Shahack-Gross, R., Ayalon, A., Goldberg, P., Goren, Y., Ofek, B., Rabinovich, R.,Hovers, E., 2008. Formation processes of cemented features in karstic cave sitesrevealed using stable oxygen abd carbon isotopic analyses: a case study atmiddle paleolithic Amud Cave, Israel. Geoarchaeology 23 (1), 43e62.
Skaberne, D., Turk, I., Turk, J., 2015. The Pleistocene clastic sediments in the Divjebabe I cave, Slovenia. Geochemical characterisation and provenance (Part 2).Palaeogeography, Palaeoclimatology, Palaeoecology 438, 379e394.
Tiessen, H., 1995. Phosphorous in the Global Environment: Transfers, Cycles andManagement. Scope 54. Wiley, New York.
Trappe, M., 2010. Actualistic approaches to the systematics and facies relations ofclastic karst deposits. Geodinamica Acta 23 (1), 41e48.
Tribovillard, N., Algeo, T.J., Lyons, T., Riboulleau, A., 2006. Trace metals as paleo-redox and paleoproductivity proxies: an update. Chemical Geology 232,12e32.
Ward, J.H., 1963. Hierarchical grouping to optimize an objective function. Journal ofthe American Statistical Association 58, 236e244.
Weiner, S., Goldberg, P., Bar-Yosef, O., 2002. Three dimensional distribution ofminerals in the sediments of Hayomin Cave, Israel: diagenetic processesand archaeological implications. Journal of Archaeological Science 29,1289e1308.
Wells, E.C., Terry, R.E., Hardin, P.J., Parnell, J.J., Houston, S.D., Jackson, M.W., 2000.Chemical analysis of ancient anthrosols at Piedras Negras, Guatemala. Journal ofArchaeological Science 27, 449e462.
White, W.B., 2007. Cave sediments and paleoclimate. Journal of Cave Karst Studies69, 76e93.
Wilson, C.A., Davidson, D.A., Cresser, M.S., 2008. Multi-element soil analysis: anassessment of its potential as an aid to archaeological interpretation. Journal ofArchaeological Science 35, 412e424.
