Human trace on the landscape during the Holocene at El Mirador Cave (Sierra de Atapuerca,Spain): The palynological evidence

The vegetation and environmental evolution

The pollen spectra identified suggest that the landscape may have been made up of a mixed forest of Pinus, deciduous and evergreen Quercus and other deciduous trees like Corylus, Fagus and Juglans. We identified a riverside community comprising Salix, Tilia, Ulmus, Alnus and hydro-hygrophilous taxa like Cyperaceae, Typha–Sparganium and the aquatic Myriophyllum. The presence of these hydro-hygrophilous taxa implies the existence of a water source near the cave on the past. At least Castanea and deciduous Quercus (e.g. Q. pyrenaica) may have lived on the acidophilous Arlanzón terraces, and some species of mountain pines and Betula must be regarded as having been transported by wind over a long distance from the higher elevations of the neighbouring regional mountain ranges, like the northwest foothills of the Iberian System (Carrión et al., 2010; Franco-Múgica et al., 2001; López-Sáez et al., 2010). It must be emphasised that the continuous presence of Castanea wich matches with its expansion from the northeast to the northwest of the Iberian Peninsula, during the middle Holocene (García-Antón et al., 1990). With respect to Abies, it should have travelled from the central and western Pyrenees, where its expansion has been documented after 6000 BP (Alba-Sánchez et al., 2009). Meanwhile, Fagus could be linked with the documented communities of the Iberian System (López-Merino et al., 2008) or the Cantabrian massifs and Pyrenees (Martínez-Atienza and Morla-Juaristi, 1992) that have acted as refuge areas until the beech expansion in the middle Holocene (Magri, 2008).

Heliophilous taxa such as Cistus or Helianthemum would have lived in the open areas within the forest. In the meadows, we would have found Poaceae, Asteraceae, Chenopodiaceae, Plantago spp., Apiaceae, Artemisia, Urticaceae, Ranunculaceae, Asphodelus and other herbs. Some of these deforested areas may have been devoted to the cultivation of cereal crops, but not continuously.

Therefore, the relatively diverse plant taxonomical composition, and its development throughout the profile, suggests the existence of a temperate climate context with some fluctuations in temperature and humidity, marked by oscillation in the representation of thermophilous and mesophilous trees and the riparian community.

The most temperate and wet phase was detected at the bottom of the sequence, where we found the highest values of arboreal taxa. This stage is consistent with the climatic stability and the thermo-mesophilous taxa expansion described in the climatic optimum (ca. 10,000 and 6500 cal. yr BP) for the Mediterranean area (Giralt et al., 1999; Jiménez-Moreno and Anderson, 2012), the north (Peñalba-Garmendia et al., 1997; Sánchez-Goñi and Hannon, 1999) and the southwest Iberian Peninsula (Fletcher et al., 2007; Reed et al., 2001).

The most important trend in the evolution of the vegetation is the progressive decrease in the arboreal pollen percentage throughout the Neolithic occupation. We observed a gradual opening up of the environment, which became much more intense as we approached the Bronze Age levels. This involved the appearance of heliophilous trees and shrubs, the development of grassland areas and the increase in vegetation associated with crops (Figure 4).

In fact, like at El Mirador, the arboreal pollen values from palynological studies conducted at archaeological sites in the north and centre of the Iberian Peninsula are generally low (Atienza-Ballano, 1995; Bueno et al., 2002; López-García and López-Sáez, 2000; López-Sáez and Burjachs, 2002; López-Sáez and López-Merino, 2007; Ruiz-Zapata et al., 2003, 2006). This is often directly related to human disturbance since AP rates are much lower than those evidenced in natural sequences for this geographical area and chronological period (López-García, 1978).

With regard to the relatively high and constant representation of pine through the sequence, it should be noted that pine is also the most represented arboreal taxa in natural and archaeological deposits available for this geographical area and chronological period (Carrión et al., 2010; Franco-Múgica et al., 2001; García-Antón et al., 2011; López-Sáez et al., 2010). In fact, the presence of pine is considered to be linked to regional vegetation and to be directly related to the surrounding mountain ranges of the Northern Plateau because of its great capacity for pollen dispersal. Pine is present from the late Pleistocene in many of the sequences of the central and coastal Iberian Peninsula (Carrión, 2002; Franco-Múgica et al., 2001, 2005; López-Sáez et al., 2014; Peñalba-Garmendia, 1994) and would have prevailed during the first half of the Holocene, with some oscillations.

This constant presence of pine during the Holocene on the Northern Plateau of the Iberian Peninsula is evident in sequences such as Cueva del Moro (López-García and López-Sáez, 1994), Gallocanta (Burjachs et al., 1996), Espinosa de Cerrato (Franco-Múgica et al., 1996) and Lillo (Franco-Múgica et al., 1997), which show the great prevalence of pine forests in the Mediterranean region. In fact, Pinus dominated, together with Quercus, Betula and Juniperus, from the Last Glacial Maximum until the influence of the establishment of deciduous trees in Atlantic areas during the Holocene (Carrión et al., 2010; Peñalba-Garmendia, 1994).

In any case, although the elevation and location of the deposits imply variations in the representation of pine and other taxa, the general trend is a fluctuation in the role of Pinus and Quercus based on small climatic variations or human pressure (Carrión et al., 2010). In fact, it seems that Pinus pinaster could be dominant in some of the regional mountain ranges and also could have been part of the mesophytic forest formations with Quercus (Carrión et al., 2000, 2007).

In the case of El Mirador, palynological data reinforce the idea of the coexistence of both types of forests during the Holocene and both regional and local presence of Pinus. The anthracological data are consistent with that coexistence throughout the sequence, but with a dominance of Quercus compared with the occasional presence of Pinus sylvestris type. The palynological data, however, illustrate a close similarity in the dynamic of the Pinus and Quercus curves. The lines only occasionally diverge, which may correspond to an aridification episode, as evidenced in Zone B1 (Figures 3 and 4). In fact, this phase related to MIR5 and the base of MIR4 level lacks evidence of anthropogenic impact on vegetation. This fact also occurs throughout the Peninsula, the central area of the Iberian Peninsula, and could match with the 4.2 ka cal. BP event (Magny, 2004; Magny et al., 2009).

So, overall, we can confirm the existence of a phase of increased aridity in the upper part of the sequence, which would coincide with the aridity episodes described for the interval 5300–3540 cal. yr BP, closely linked to the expansion of pine, sometimes in conjunction with evergreen Quercus (Carrión et al., 2010). In El Mirador, the beginning of this phase coincides with level MIR5 (chronologically between 3560 and 5600 cal. yr BP), and the rise of Pinus starts becoming irrefutable in MIR4. It is precisely in this level, which belongs to the Bronze Age phase, where we observed the maximum representation of pine and an increase in the Mediterranean trends and aridity characteristic of the beginning of the middle Holocene in this geographic area and that would have influenced the decline or disappearance of Ulmus, Juglans, Tilia and Fagus populations (Carrión et al., 2010; Franco-Múgica et al., 2001, 2005; García-Antón et al., 2011; Gil-García et al., 2002; López-Sáez et al., 2003; Muñoz-Sobrino et al., 2004). The incidence of aridity and the intensification of human pressure are also reflected in the anthracological record, which shows a change in dominance from deciduous Quercus to evergreen. This change is noticeable from MIR14 (6320–6200 cal. yr BP) and is definitely consolidated at the top of the sequence (Allué and Euba, 2008).

It is difficult to document the influence of these xerophytic phases in the settlements in the vicinity of the Sierra de Atapuerca, although some evidence of the occurrence of some of these aridity episodes has been found. This is the case of the 8.2 ka BP event, verified from the pollen record of Cañada de la Cruz, which caused the migration of most of the hunter–gatherer communities from the central Ebro basin (González-Sampériz et al., 2009). In our case, it was not possible to assess the impact of natural forest fires so closely related to processes of aridity. The genesis of the deposit does not allow us to differentiate between charcoals from fires and those generated by anthropic systematic burning. At any rate, the incidence of decreased rainfall is parallel to the progressive increase in indicators of anthropic pressure. Thus, the oscillation of the representation of arboreal pollen at El Mirador is clearly accompanied by the appearance of vegetation typical of a deforestation phase, the development of anthropozoogenous grasslands and the proliferation of nitrophilous or ruderal plants (Barbier et al., 2001; Behre, 1981; López-Sáez et al., 1998).

Human activities and anthropogenic modifications

Although one of the objectives of this work has been the landscape reconstruction in the vicinity of the Sierra de Atapuerca during the Neolithic and Bronze Age, the most important trend identified in the sequence analysed is the role of vegetation related to a clear anthropised environment.

It should be noted that in the data from bogs and palaeolakes of the north and central areas of the Peninsula, such as Comeya (Jiménez-Sánchez et al., 2003), Monte Areo and Alto de la Espina (López-Merino, 2009), Puerto de los Tornos (Peñalba-Garmendia, 1994), Turbera de la Piedra (Muñoz-Sobrino et al., 1996), Quintanar de la Sierra (Peñalba-Garmendia, 1989), Hoyos de Iregua (Gil-García et al., 2002), Lago del Carrizal (Franco-Múgica et al., 2005), Espinosa de Cerrato (Franco-Múgica et al., 2001) and Cañada de la Cruz (González-Sampériz et al., 2009), the first pollinical evidence of the start of anthropisation appears only in the early Neolithic.

Nevertheless, the first clear signs of human impact are not found in the Iberian System until the late Bronze Age (Ruiz-Zapata et al., 2002), when the higher elevation areas of the Northern Plateau began being colonised. This late sign of human communities’ influence on the landscape is surely related to the elevation at which these deposits are found. It seems obvious that the first more or less permanent settlements that exerted pressure on the environment were located in the lowlands and valleys, so a pollen analysis that only provides a regional view of the landscape would not detect these signs (Carrión et al., 2010). This increase in the phenomenon of anthropic activity has been widely observed in most palynological sequences available for the north and the central areas of the Iberian Peninsula at sites such as Can Roqueta (Burjachs and Expósito, 2007), Lago Enol (López-Merino, 2009), Rascafría (Ruiz-Zapata et al., 2006), El Portalón (Ruiz-Zapata et al., 2003) and Covatilla (Atienza-Ballano, 1995).

The palynological analyses of archaeological deposits such as Azutan (Bueno et al., 2002), Barruecos (López-Sáez et al., 2007), Cueva de Chaves (López-Sáez et al., 2006), La Cueva de los Gitanos (Ontañón et al., 2013) and Dehesa Río Fortes (López-Sáez, 2002) reflect an evident human modification of the vegetation on the north and centre of the Iberian Peninsula at the end of the Atlantic period (Carrión et al., 2010; Delibes de Castro et al., 2015; López-Sáez et al., 2014) as in the rest of European continent (Zvelebil, 1989).

The first archaeobotanical evidence of agricultural practices on the Iberian Peninsula was documented on the east coast, at Mas d’Is, Valencia, dating to ca. 6600 ± 50 BP (Bernabeu-Aubán et al., 2003). The available data from inner Iberia have documented the agriculture practices in some sites in the Meseta as La Lámpara and La Revilla (Rojo-Guerra et al., 2006) and Cueva de la Vaquera (Estremera, 2003), among others, around 6400 cal. yr BP. In the northeast of the Iberian Peninsula, Can Sadurní (Blasco et al., 2005) has reported evidences of farming activities at 6450 ± 50 BP, and in the rockshelter of Chaves, located in the Ebro valley, cereal pollen have been identified between 6230 and 6120 BP (López-Sáez and López-García, 2000).

In this way, Neolithisation process implies significant changes in the relation between environment and human communities. These transformations spread throughout the Iberian Peninsula at various stages of inland colonisation through the channels of major rivers of the Mediterranean coastal area (Bernabeu-Aubán et al., 2014). This process starts in the first half of the eighth millennium completing the cultural and social change that had begun in the southeast of the European continent in the ninth millennium (Marinova et al., 2012; Revelles, 2016).

The palynological sequence of El Mirador reflects reliably the introduction and settlement of this new production system characterised, in the Iberian Peninsula, for the great diversity of the crops including a large variability of cereals and vegetables, from ca. 7970 to 7770 cal. yr BP (Zapata et al., 2004).

In El Mirador, the agricultural practices have already been demonstrated through the data provided by the archaeobotanical analysis (Rodríguez-Cruz and Buxó, 2008). The presence of seeds of Triticum aestivum/durum and Triticum dicoccum was identified from MIR6 to MIR24, as well as, exceptionally, Hordeum vulgare var. nudum. Agricultural by-products have also been identified in 10 levels, although with a greater presence at the base of the sequence (from MIR19 to MIR24).

The palynological analysis revealed that the Cerealia-type curve is continuous throughout the entire sequence with some fluctuations, perhaps related to the nature of the activities carried out by the communities who occupied the cave. Indeed, there seems to be an alternation between livestock and farming practices that also involves changes in the values of the nitrophilous vegetation component, in accordance with the intensity of the occupations. Still, it is necessary to thoroughly assess the significance of the oscillation of the different indicators. For example, at the top of the sequence, the decline in nitrophilous plants, coinciding with a rise in cultivation, does not appear to be exclusively related to a lesser degree of anthropisation, but rather to a change in land use strategies. According to López-Sáez et al. (2000), areas previously occupied by nitrophilous and ruderal plants were afterwards used for crops, and this could have been the case at El Mirador.

Nevertheless, in some cases, it has identified a very intensive development of livestock. Thus, the progressive decrease in Poaceae observed in some moments of the sequence was probably related to the high incidence of farming practices in an arid or less humid environment, which may have been decisive in the decrease in grassland areas near the cave at the beginning of the Bronze Age. So the increase in percentages of pastoral pressure indicators that match with the minimal representation of Poaceae could be related to the depletion of these types of herbaceous plants by overgrazing (Burjachs et al., 1997) and could involve the existence of a more sedentary mobility pattern in these communities.

In any case, the Cerealia percentage exceeds 3% in 27 of the samples analysed, which may be related to the existence of crop fields near the cave (Diot, 1992; Richard, 1995). Values close to 10%, identified in some of the levels of the sequence (Figure 4), have been attributed, however, to the existence of some degree of anthropogenic manipulation (Diot, 1992) or to an indirect contribution of anthropogenic origin (Bower, 1998; Diot, 1992; Robinson and Hubbard, 1977).

The impact on the environment intensified in the Bronze Age levels is evidenced by an increase in anthropogenic indicators. The exceptional location of the settlement probably made it possible to establish cultivation areas in the vicinity and to use large areas of grassland. In addition, the strategic location of El Mirador cave, connecting two large valleys, suggests that this site that would have been frequented by pastoralist groups of the Northern Plateau moving to the area of Cantabrian influence or the Iberian System to reach high-mountain pastures.

The only point in which there is no evidence of these activities at El Mirador is in MIR5 which, as mentioned above, represents a unique and uncommon phenomenon. Thus, palynologically, this level yielded evidence of plants and NPPs opposite to those associated with anthropogenic contexts, as well as the existence of a particularly dry climate context.

Regarding the contribution of NPPs and considering the functionality of the infilling and the fact that it was subjected to periodic burning for the purpose of cleaning and deworming, one would expect greater taxonomic variability and a greater percentage of NPPs, and especially coprophilous and carbonicolous ones. The low presence of these kinds of microorganisms leads us to consider that the systematic fires may have inhibited the activity of these saprophytic organisms.

The oscillation in the representation of taxa associated with nitrogenated contexts (Figure 6) indicates that the changes in the trophic conditions can be explained by the increase or decrease in phosphate- and nitrate-rich waste related to pastoral pressure (Van Geel et al., 1981). Naturally, the systematic burning of infill drastically reduces the eutrophic content of the soil, which is regenerated through the proliferation of species like Gloeotrichia, a pioneer alga that colonises nitrogen-poor substrates (Van Geel, 1986; Van Geel et al., 1984, 1989), or through a new occupation of the cave.

OPEN IN VIEWER

Figure 6.

Synthetic palynological diagram of El Mirador Cave. The groups are formed according to the taxa allocation shown in Table 2.