Scholarly article on topic 'Reconstructing the impact of human activities in a NW Iberian Roman mining landscape for the last 2500 years'

Reconstructing the impact of human activities in a NW Iberian Roman mining landscape for the last 2500 years Academic research paper on "History and archaeology"

CC BY-NC-ND
0
0
Share paper
Academic journal
Journal of Archaeological Science
OECD Field of science
Keywords
{"Pollen analysis" / "Forest clearance" / Mining/metallurgy / "Human impact" / "Forest resilience"}

Abstract of research paper on History and archaeology, author of scientific article — Lourdes López-Merino, Antonio Martínez Cortizas, Guillermo S. Reher, José A. López-Sáez, Tim M. Mighall, et al.

Abstract Little is known about the impact of human activities during Roman times on NW Iberian mining landscapes beyond the geomorphological transformations brought about by the use of hydraulic power for gold extraction. We present the high-resolution pollen record of La Molina mire, located in an area intensely used for gold mining (Asturias, NW Spain), combined with other proxy data from the same peat core to identify different human activities, evaluate the strategies followed for the management of the resources and describe the landscape response to human disturbances. We reconstructed the timing and synchronicity of landscape changes of varying intensity and form occurred before, during and after Roman times. An open landscape was prevalent during the local Late Iron Age, a period of relatively environmental stability. During the Early Roman Empire more significant vegetation shifts took place, reflected by changes in both forest (Corylus and Quercus) and heathland cover, as mining/metallurgy peaked and grazing and cultivation increased. In the Late Roman Empire, the influence of mining/metallurgy on landscape change started to disappear. This decoupling was further consolidated in the Germanic period (i.e., Visigothic and Sueve domination of the region), with a sharp decrease in mining/metallurgy but continued grazing. Although human impact was intense in some periods, mostly during the Early Roman Empire, forest regeneration occurred afterwards: clearances were local and short-lived. However, the Roman mining landscape turned into an agrarian one at the onset of the Middle Ages, characterized by a profound deforestation at a regional level due to a myriad of human activities that resulted in an irreversible openness of the landscape.

Academic research paper on topic "Reconstructing the impact of human activities in a NW Iberian Roman mining landscape for the last 2500 years"

ELSEVIER

Contents lists available at ScienceDirect

Journal of Archaeological Science

journal homepage: http://www.elsevier.com/locate/jas

Reconstructing the impact of human activities in a NW Iberian Roman mining landscape for the last 2500 years

Lourdes López-Merino a *, Antonio Martínez Cortizas b, Guillermo S. Reher c, José A. López-Sóez d, Tim M. Mighall e, Richard Bindlerf

a Institute for the Environment, Brunei University, Uxbridge, London, Middlesex UB8 3PH, UK

b Departamento de Edafología y Química Agrícola, Facultad de Biología, Universidad de Santiago, Campus Sur, 15782 Santiago de Compostela, Spain c G.I. Estructura Social y Territorio — Arqueología del paisaje, Instituto de Historia (CCHS), Consejo Superior de Investigaciones Científicas (CSIC), Albasanz 26-28, 28037 Madrid, Spain

d G.I. Arqueobiología, Instituto de Historia (CCHS), Consejo Superior de Investigaciones Científicas (CSIC), Albasanz 26-28, 28037 Madrid, Spain e Department of Geography and Environment, School of Geosciences, University of Aberdeen, Elphinstone Road, Aberdeen AB24 3UF, UK f Department of Ecology and Environmental Science, Umea University, SE-901 87 Umea, Sweden

CrossMark

ARTICLE INFO

Article history: Received 31 March 2014 Received in revised form 9 June 2014 Accepted 10 July 2014 Available online 23 July 2014

Keywords: Pollen analysis Forest clearance Mining/metallurgy Human impact Forest resilience

ABSTRACT

Little is known about the impact of human activities during Roman times on NW Iberian mining landscapes beyond the geomorphological transformations brought about by the use of hydraulic power for gold extraction. We present the high-resolution pollen record of La Molina mire, located in an area intensely used for gold mining (Asturias, NW Spain), combined with other proxy data from the same peat core to identify different human activities, evaluate the strategies followed for the management of the resources and describe the landscape response to human disturbances. We reconstructed the timing and synchronicity of landscape changes of varying intensity and form occurred before, during and after Roman times. An open landscape was prevalent during the local Late Iron Age, a period of relatively environmental stability. During the Early Roman Empire more significant vegetation shifts took place, reflected by changes in both forest (Corylus and Quercus) and heathland cover, as mining/metallurgy peaked and grazing and cultivation increased. In the Late Roman Empire, the influence of mining/metallurgy on landscape change started to disappear. This decoupling was further consolidated in the Germanic period (i.e., Visigothic and Sueve domination of the region), with a sharp decrease in mining/ metallurgy but continued grazing. Although human impact was intense in some periods, mostly during the Early Roman Empire, forest regeneration occurred afterwards: clearances were local and short-lived. However, the Roman mining landscape turned into an agrarian one at the onset of the Middle Ages, characterized by a profound deforestation at a regional level due to a myriad of human activities that resulted in an irreversible openness of the landscape.

© 2014 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND

license (http://creativecommons.org/licenses/by-nc-nd/3.0/).

1. Introduction

The Roman period is well known for its large-scale impact on the landscapes and environment of central and southern Europe. It ranged from local imprints left by agriculture, animal husbandry and mining, to regional and even hemispheric-scale atmospheric

* Corresponding author. Tel.: +44(0)1895 267311; fax: +44(0)1895 269761. E-mail addresses: lolome@hotmail.es, lourdes.lopez-merino@brunel.ac.uk (L López-Merino), antonio.martinez.cortizas@usc.es (A. Martínez Cortizas), guillermo. reher@cchs.csic.es (G.S. Reher), joseantonio.lopez@cchs.csic.es (J.A López-Sóez), t. mighall@abdn.ac.uk (T.M. Mighall), richard.bindler@emg.umu.se (R. Bindler).

metal pollution due to mining and metallurgy, which was detected in Greenland ice cores (Hong et al., 1994). Such a profound change in the scale of resource exploitation caused a restructuring of landscape management practices that was even more significant in mining areas, where the landscape was completely reorganised with new settlement patterns as a consequence of deep transformations of the territorial structures. One of the regions most affected was the north-western part of the Iberian Peninsula, due to its extensive gold resources (Domergue, 1987). The Conventus Iur-idicus Asturum (mainly modern León and Asturias provinces in NW Spain) was rich in gold deposits, a strategic resource crucial for maintaining the Imperial coinage system, which was sustained by

http://dx.doi.org/10.1016/jjas.2014.07.016

0305-4403/© 2014 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).

the aureus. The Roman extractive activities, which involved the harnessing and use of waterpower, triggered important geomor-phological changes that resulted in the shaping of a completely new landscape, which is still clearly visible today. One of the most famous examples in NW Spain is found in Las Módulas (León) (Orejas and Sónchez-Palencia, 2002), a cultural landscape that was declared as an UNESCO World Heritage Site. Archaeological and archaeobotanical research has shown the various ways in which pre-Roman and Roman communities managed their surroundings through different modes of production (Saónchez-Palencia, 2000; López-Merino et al., 2008, 2010a; Sanchez-Palencia et al., 2011).

The chronology of the Roman gold mining activities in this area is not known precisely but it is generally accepted that it started after the conquest and territorialisation carried out under Augustus (27 BC—AD 14) and intensified in the Flavian period (late 1st century AD). The end of this imperial model of exploitation apparently took place towards the beginning of the 3rd century, at the end of the Early Roman Empire (Edmondson, 1989). Knowledge of Roman gold mining in Iberia has expanded greatly in recent decades, stemming particularly from the prolific work of Claude Domergue, who specialized in Hispania as the richest gold-producing region in the Ancient World (Domergue, 1987). In the León province, the techniques used (Domergue and Hórail, 1978), the working conditions (Orejas, 1994), the implications mining had for social processes (Orejas and Sanchez-Palencia, 2002), and even the environmental impact (Loópez-Merino et al., 2010a) are now much better understood. In contrast, Roman gold mining in Asturias province (Fig. 1) has only been systematically studied using a landscape archaeology approach by Sónchez-Palencia (1983), leaving two published general syntheses that enumerate gold mining throughout the western half of the province (Sanchez-Palencia and Suórez, 1985; Sanchez-Palencia, 1995). Little is known about the impact of mining and other human activities on

the landscapes of Asturias beyond the geomorphological changes brought about by hydraulic power.

One of the first indications of regional metalworking in the palaeoenvironmental record is the detection of atmospheric metal pollution. Although metal pollution linked to metallurgical activities has occurred for more than three millennia in NW Iberia (Pontevedra-Pombal et al., 2013), it climaxed during Roman times (Martínez Cortizas et al., 1997, 1999, 2002, 2013; Kylander et al., 2005). However, the impact of mining and metallurgical activities was not restricted only to metal pollution or the above-mentioned geomorphological changes. Studies of other historical mining areas of Europe (e.g., Mighall and Chambers, 1997; Monna et al., 2004a, 2004b; Jouffroy-Bapicot et al., 2007; Breitenlechner et al., 2010; Curras et al., 2012) also found ecological impacts such as forest clearance to promote agriculture and/or animal husbandry, and the use of wood resources for mining and metalworking, including charcoal production.

Recognizing the impact of past human activities, e.g., resource exploitation and management, on the environment and landscapes is essential for unravelling the origin, history and trajectories of current continental ecosystems (Birks, 2012), and key for the understanding of their response to a range of perturbations and their resilient behaviour (Dearing, 2006; Dearing et al., 2006). In this sense, multi-proxy palaeoecological studies have an important role in identifying and assessing the multiple intertwined forces that operate as long-term processes that control environmental transformation. We present a record of the last c. 2500 years of landscape change through the pollen study of a peat core sampled in La Molina (Asturias, Spain; Fig. 1); a mire located in an area where extensive gold mining and metallurgical activities took place during Roman times. As already mentioned, little is known about the impact of past human activities other than the geomorphological changes produced by mining activities. In order to redress this research gap and develop a complete picture of the complexity of

Fig. 1. A) and B) Location of La Molina mire in NW Iberia; C) local mining activity in Roman times (Perea and Sanchez-Palencia, 1995).

human intervention on this landscape, we synthesize the new pollen data with previous studies performed on the same peat core. The specific aims of this study were i) to identify different human activities, ii) to evaluate the strategies followed for the management of the resources at different cultural periods, and iii) to describe the landscape response to the human disturbances.

2. Setting and research background

2.1. Environmental and archaeological settings

La Molina is a Site of Community Importance (SCI) in the Natura 2000 network due to its ombrotrophic nature (habitat 7110 Raised bogs). The mire covers a surface of 32.8 ha and is located in the Alto de la Espina Range at 650 m a.s.l. (Concejo de Salas, Asturias, Spain; 43o 22' 52.2" N, 6o 19' 38.4" W; Fig. 1). The mire complex is represented by acidic, Sphagnum-dominated mesotopes, with the presence of Sphagnum pylaisii. The remaining forests in the area

include oak (Quercus robur), hazel (Corylus avellana), beech (Fagus sylvatica), birch (Betula alba), and sweet chestnut (Castanea sativa). Most of the original area of forest has been replaced by pine and eucalyptus afforestation and heathlands dominated by Erica species, Calluna vulgaris and other shrubs (Ulex and other Genisteae). The use of some hillsides for agriculture and grazing has expanded the area covered by cereal crops and pasture.

The mire is close to a complex comprising four mines known as Ablaneda (Domergue, 1987), 5.5 km to the east (Fig. 2). These have been known since the mid-19th century, as well as the parallel canals that led to them at slightly different altitudes, hugging the quartzite cliff of Peñausende (Martínez Alcíbar, 1850; Schulz and Paillette, 1850; Paillette, 1852). Accidental breaching of one of these channels revealed a structure that closely resembled other examples from neighbouring areas such as Las Meódulas (Saónchez-Palencia, 2000), merely 1 m wide and 30 cm deep (Camino Mayor, 1989a). A more recent study has offered a further overview of the mining complex, from the water reservoir at La Molina to the mines

Fig. 2. Mining complex of Ablaneda, including the mines, the canals and the deposit at La Molina (Las Mueyes), as interpreted from high-resolution satellite imagery. Associated microtoponyms mentioned in the text are shown in italics. Settlements belong, surely or presumably -given the scarce data available and other regional morphologies-, to the Roman period, when the mine was active. The road is the itinerary of the medieval communication route which could presumably have existed also previously.

at Ablaneda (Fanjul Peraza and Menéndez Bueyes, 2003—2007). The name La Molina, which refers to local mills, is alternative to the local place name usually bestowed on this area — Las Mueyes/Las Mueches. The remains of a large stone wall, which closed off the reservoir, the Murón (Fig. 2), can barely be seen today (Camino Mayor, 1989b), because the stones have been reused. Measurements taken from the top of the wall to the lowest part of the mire basin indicate a depth of 4.5—5.0 m, making this the largest Roman water reservoir known in NW Iberia. Several channels leave the reservoir. They were probably used during different phases of mining and ran to various destinations within the mine (Fig. 2). Other channels also lead to the same area, but stem from another local stream (Fanjul Peraza and Meneóndez Bueyes, 2003—2007).

2.2. Previous palaeoenvironmental research

López-Merino et al. (2011) studied the complex pattern of hydrological shifts in the mire during the last c. 2500 years using the records of hydro-hygrophytes and non-pollen palynomorphs (NPP) together with some geochemical data. This research showed that the mire was minerogenic in the local Late Iron Age (minerotrophic phase), but since Roman times it was subjected to hydrological changes due to a rise in the water-table, fluctuating between the presence of open water and phases of drawdown (transitional phase) (Fig. 3). The changes were most likely due to the direct use of the wetland as a water-reservoir for the canalisation system used for gold-mining. In post-Roman times it gradually evolved towards ombrotrophy (ombrotrophic phase), with increased grazing pressure more recently. Additionally, Martínez Cortizas et al. (2013) studied the same peat core using major, minor and trace lithogenic elements, trace metals/metalloids and stable Pb isotopes. This study indicated that La Molina preserved a very detailed record of atmospheric lead pollution and catchment soil erosion (i.e., mineral inputs to the mire) during

the last c. 2500 years (Fig. 3). Atmospheric metal pollution follows an increasing trend during the Roman Period with several peaks, while the mineral inputs to the mire decreased showing also several phases. The detected seesaw pattern suggests significant variability at short-time scales — probably related to the local history of mining in NW Iberia.

3. Material and methods

3.1. Sampling and chronology

A 215 cm-deep core was collected in July 2005 using a Russian peat corer (50 cm long and 5 cm in diameter). Peat sections were placed in PVC tubes, protected in plastic guttering and stored under cold conditions (4 °C) prior to laboratory sub-sampling and analysis. The core was sectioned into continuous 1 cm-thick slices, except for the upper 4 cm of living vegetation. In this article we present data for the top 115 cm. For the whole sequence see López-Merino (2009). The core stratigraphy consists of four parts: I (115—87 cm), brown peat rich in mineral matter; II (87—43 cm), black peat with four layers rich in mineral matter (78—72, 69—65, 61—57 and 51—49 cm); III (43—4 cm), brown moss peat with decomposed Sphagnum remains; and IV (4 cm-top), the living vegetation.

The age—depth model, constructed with the aid of six AMS radiocarbon dates, was published by López-Merino et al. (2011), and constrained by Martínez Cortizas et al. (2013) assigning an age of AD 1975 to the near-surface peat which shows the largest Pb concentration and the recent lowest 206Pb/207Pb ratio as found in previous investigations in NW Spain (Kylander et al., 2005; Martínez Cortizas et al., 2012). The model was obtained using the Clam application developed by Blaauw (2010). The best fit was obtained with a fourth order polynomial (Supplementary Information). According to the model, the base of the 115 cm

Fig. 3. Compilation of the main results obtained in previous studies on La Molina mire for the last c. 2500 years. A) Mineral content of the peat, a proxy for catchment soils erosion (PC1 factor scores of the geochemical data; Martínez Cortizas et al., 2013); B) atmospheric metal pollution index, proxy for mining and metallurgy (PC2 factor scores of the geochemical data, note that the X axis has been inverted, Martínez Cortizas et al., 2013); C) local wetland phases (López-Merino et al., 2011); D) water-table level index (sum of PC1, PC7, PC4, PC6, PC3 and PC8 factor scores of the hydro-hygrophytes and NPP data, as they are related to fluctuations in the water-table levels, data from López-Merino et al., 2011); E) grazing index (PC2 factor scores of the hydro-hygrophytes and NPP data, related to coprophilous fungi, López-Merino et al., 2011).

Fig. 4. Percentage pollen diagram of La Molina mire (woody vegetation) plotted against age. The filled silhouettes show the percentage curves of the taxa, the white silhouettes show the x 10 exaggeration curves. Dots represent percentages below 0.5%.

section dates to the local Late Iron Age (c. 2500 cal yr BP) and the sequence extends up to the present day without any sedimentary hiatus recorded.

3.2. Pollen analysis

Palynomorphs were isolated following the classic procedure (Moore et al., 1991) with concentration in heavy liquid. Pollen counting was conducted at 400 x magnification, and a minimum of 400 terrestrial pollen grains were considered in the pollen sum, the lowest pollen sum being 435. Pollen identification was aided by the reference collection of the Laboratory of Archaeobiology at the CCHS (Madrid), standard identification keys (Moore et al., 1991) and the photographic atlas (Reille, 1992) of European pollen flora.

Additionally, Pinus pinaster and Quercus suber types were identified following the criteria provided by Carrion et al. (2000a, 2000b). The analysis was conducted in 101 samples taken every 1 (from Roman times onwards) or 2 (late Iron Age) cm. The core has a particularly good resolution for the Roman period, with 40 pollen samples spanning c. 20 BC to AD 440, providing nearly decadal resolution. The pollen and the summary diagrams (Figs. 4—6) have been plotted against age using Tilia. The pollen sequence was divided into five pollen zones (Pz-1 to Pz-5) according to the local cultural periods. Palynological richness was estimated by rarefaction analysis using the open software PAST 3.01 (Hammer et al., 2001). The number of grains selected for standardization in the rarefied sample was the lowest pollen sum (435).

Fig. 5. Percentage pollen diagram of La Molina mire (herbs) plotted against age. The filled silhouettes show the percentage curves of the taxa, the white silhouettes show the x10 exaggeration curves. Dots represent percentages below 0.5%. Palynological richness calculated by rarefaction analysis is also shown, as well as the total land pollen sum.

Previous data have been used in this study to develop a more complete picture of the environmental and landscape changes since the local Late Iron Age. Firstly, the grazing index obtained by López-Merino et al. (2011) after a principal component analysis (PCA) applied on the hydro-hygrophytes and non-pollen paly-nomorphs record was adopted. This principal component is associated to positive loadings of Cercophora-type, Chaetomium, Podospora-type, Sordaria-type and Sporormiella-type. These NPP are ascospores of coprophilous fungi, indicating herbivore pressure on the mire, and their curves follow a comparable pattern with the one exhibited by the principal component scores. Secondly, the atmospheric metal pollution index obtained by Martínez Cortizas et al. (2013) after a PCA on geochemical data has also been used. The extracted principal component is mainly related to changes in the 206Pb/207Pb ratio, indicating metal (Pb) pollution. The correlation between these indexes with selected pollen types (deciduous Quercus, Corylus and Erica t.) was assessed using Pearson's correlation coefficients, for different cultural phases from the Late Iron Age to the Germanic period (Table 1). Metal pollution factor scores were first multiplied by -1 for the calculation of Pearson's correlation coefficients and for Fig. , so the larger the positive score of a sample the higher the metal pollution is. This was done to make the interpretation of the results more intuitive.

4. Results

4.1. Pz-1: Late Iron Age (115—90 cm; c. 500—20 BC)

The arboreal pollen component is dominated by Corylus and deciduous Quercus, with the continuous presence of Alnus and Betula. Other mesophytes such as Fraxinus, Fagus, Castanea, Salix and Ulmus are also present, although with low values. Shrubs are abundant, with Erica t., C. vulgaris and Rosa t. being the most important taxa (Fig. 4). Poaceae is the main herbaceous component. Plantago lanceolata t. and Plantago major/media t. are present, as well as other herbs such as Liliaceae, Fabaceae, Rumex acetosella t., Cardueae and Cichorioideae. Cerealia t. presents low values and a discontinuous record (Fig. 5).

4.2. Pz-2: Roman times (90—50 cm; c. 20 BC—AD 440)

Deciduous Quercus and Corylus are still the most abundant trees, and other mesophilous trees such as Fraxinus, Ulmus and Acer have higher percentages than previously recorded. Additionally, cultivated and mesothermophilous taxa such as Castanea, Juglans, evergreen Quercus, Q suber, Arbutus, Olea europaea and P. pinaster are present, but with low values (Fig. 4). Although tree percentages are relatively important, deciduous Quercus and Corylus values show a seesaw pattern (Fig. 6). Amongst the shrub component, Erica t., C. vulgaris and Rosa t. values are lower than those previously detected (Fig. 4). Herbaceous types such as Urtica dioica t., R. acetosella t., Artemisia and Fabaceae occur in higher values than in Pz-1 (Fig. 5).

Some differences between the Early Roman Empire (Pz-2a, c. 20 BC-AD 240) and the Late Roman Empire (Pz-2b, AD c. 240—440) are noticeable. During Pz-2a, Frangula, Cornus sanguinea t., Crataegus t., Rubus t. and Cerealia t. present relatively high percentages (Figs. 4 and 5). During Pz-2b, Cerealia t. is less abundant, as well as the aforementioned shrubs. Rosa t. values decrease and Cytisus/Ulex t. values increase.

4.3. Pz-3: Germanic period (50—43 cm; AD c. 440—700)

Deciduous Quercus is the most abundant tree pollen type. Cor-ylus and Fraxinus decrease in value, while Betula, Fagus and Casta-nea increase. Other tree species maintain low percentages, i.e.,

Alnus, Acer. Shrubs are not abundant, as during Pz-2, although Cytisus/Ulex t. present higher percentages than previously recorded. Erica t., C. vulgaris occur in very low values (Fig. 4). Fabaceae and Artemisia show increased representation among the herbaceous component. Cerealia t. is almost absent in this pollen zone (Fig. 5).

4.4. Pz-4: Middle Ages (43-32 cm; AD c. 700-1450)

Tree cover gradually diminishes through this zone, as indicated by the lower percentages of deciduous Quercus, Betula and Ulmus. However, Castanea percentages increase again (Fig. 4). The most notable feature is the spread of shrubs. As in Pz-3, Cytisus/Ulex t. has significant values, while Erica t. and, to a lesser extent, C. vulgaris percentages increase. Prunus t. appears for the first time (Fig. 4). Percentages of herbs such as Poaceae, Artemisia, Aster t., Cardueae, Cichorioideae, Liliaceae, Plantago lanceolata t. and Plantago major/ media t. all increase. Cerealia t. appears constantly during this zone (Fig. 5).

4.5. Pz-5: Modern Era (<32 cm; AD >1450)

Tree percentages are the lowest they have been throughout the whole sequence, while Erica t. shows high values, and other shrubs such as C. vulgaris, Cytisus/Ulex t. and Prunus t are also present (Fig. 4). Some differences exist between the Early Modern (Pz-5a, AD c. 1450-1850) and Late Modern (Pz-5b, AD >1850) periods. During Pz-5a, tree percentages are low and the values of Poaceae, Liliaceae, Plantago sp., Cardueae and R. acetosella t. are important. Cerealia t. is also present (Fig. 5). However, during Pz-5b Erica t. values show a decreasing trend, while P. pinaster and herbs (i.e., Plantago lanceolata t., Plantago major/media t., Plantago coronopus t., U. dioica t.) percentages increase, along with Cerealia t. and the appearance of Eucalyptus (Figs. 4 and 5).

4.6. Palynological richness

The results of the rarefaction analysis show a varied distribution along the record (Fig. 5). The average is 28.08 pollen types, although values range from 19.90 to 35.79 pollen types. During Pz-1, paly-nological richness values are usually lower than the average. However, a seesaw pattern begins to be evident from c. 300 BC onwards. During Pz-2a, this pattern is more evident and palyno-logical richness values show an increasing trend from the average values to 35.79 pollen types. Palynological richness decreases unevenly during Pz-2b, reaching average or lower than the average values in Pz-3, and they are more or less constant during Pz-4 and Pz-5a. Elevated palynological richness is detected in Pz-5b.

5. Discussion

The multi-proxy approach applied here enable us to reconstruct landscape change since the local Late Iron Age and to recognise the role played by different human activities, such as mining/metallurgy, cultivation, animal husbandry and deforestation. Two main periods could be distinguished: the first one characterised by the reversible effects of the human activities on the woodlands, the second one a non-regeneration of the tree cover.

5.1. Reversible effects of human impacts on the vegetation from the Late Iron Age to Germanic times

During the local Late Iron Age (Pz-1) hazel and oak were the most important components of medium altitude woodlands surroundings the La Molina mire. Additionally, shrubs such as Erica t. and C. vulgaris, together with Rosa t. and grasses, indicate an

Table 1

Pearson's correlation coefficients of selected pollen taxa, metal pollution and grazing for the Late Iron Age, Early and Late Roman Empire and Germanic Period. Note that metal pollution values haven been multiplied by -1, so the larger the score of a sample, the higher the metal pollution is. This has been done to make the interpretation of the results more intuitive.

Fig. 6. Chronology of the environmental changes detected in La Molina mire from c. 500 BC to AD 700 (from top to bottom): Metal pollution index (note that the X axis has been inverted, Martínez Cortizas et al., 2013); Corylus percentages; Erica t. percentages; grazing index (López-Merino et al., 2011); deciduous Quercus percentages; Cerealia t. percentages; and palynological richness.

already open landscape, probably due to local human activities. Despite the relative landscape stability during the first half of Pz-1, two small-scale clearance events were detected (Fig. 6). The first one (c. 300—160 BC) supposed a minor decrease in deciduous Quercus pollen, suggesting a small opening of oak woodlands surrounding La Molina, due to mining/metallurgical activities and agricultural and pastoral farming, as indicated by small peaks in atmospheric metal pollution, Cerealia t. pollen and grazing index. The second event (c. 120—30 BC) had a more sizeable impact on forest cover, although it was still a small-scale, local woodland clearance. Increases in the values of Cerealia t. and grazing index are linked to this clearance that produced a decrease of, first, hazel and, second, oak. Both clearances could be related to the development of larger, more complex settlements at the end of the Iron Age (Marín Suórez, 2011).

The overall picture is more complex during the Roman period (Pz-2), and not only in terms of human activities. The presence of mesothermophilous taxa such as evergreen Quercus, Q. suber, Arbutus and Olea europaea increased, probably due to the warmer temperatures characterising the Roman Warm Period. A reconstructed temperature index from Penido Vello (PVO), a mountain blanket bog located 100 km west of La Molina, shows warmer conditions in NW Spain at this time (Martínez Cortizas et al., 1999). A shift in the hydrology of the mire can be detected at the beginning of the Early Roman Empire (Pz-2a, c. 20 BC—AD 240), representing a large perturbation in the mire system, which triggered a rapid decrease in the mineral matter flux from the mire's catchment

Cultural period Taxa Metal pollution Grazing

Germanic period Deciduous Quercus -0.47 0.21

(AD c. 440-700) Corylus 0.05 -0.65

Erica t. -0.41 -0.42

Late Roman Empire Deciduous Quercus -0.04 -0.23

(AD c. 240-440) Corylus 0.34 -0.37

Erica t. -0.09 0.27

Early Roman Empire Deciduous Quercus 0.51a 0.42b

(c. 20 BC-AD 240) Corylus 0.69a 0.46b

Erica t. 0.76a 0.54a

Late Iron Age Deciduous Quercus -0.14 -0.06

(c. 500-20 BC) Corylus 0.01 0.03

Erica t. 0.49b 0.14

a Correlation is significant at the 0.01 level (2-tailed). b Correlation is significant at the 0.05 level (2-tailed).

(López-Merino et al., 2011). Four synchronous phases with increases in atmospheric metal pollution are distinguished, following a marked seesaw pattern as that shown by the water-table level and geochemical index of soil erosion (Martínez Cortizas et al., 2013) (Fig. 3). Additionally, Frangula, C. sanguinea t., Crataegus t. and Rubus t. are well represented (Fig. 4), probably linked to a higher wetland water-table level as they are commonly found on river banks and in disturbed areas. The above-mentioned evidence combined with the archaeological record of a channel network in the surroundings and the remains of a dam (Camino Mayor, 1989a, 1989b; Fanjul Peraza and Menendez Bueyes, 2003—2007; Fernandez Mier, 1999) (Fig. 2), strongly suggest that the mire was intentionally flooded — leading to an increase in bioproductivity and thus an increase in the organic matter content of the peat. This human perturbation was linked to mining activities, because water was necessary to open lode deposits and to wash the ore down sluices, where gold particles could be separated from the remaining ore. This was achieved by hydraulic force obtained using water-canalisation systems (corrugi) and water deposits (piscinae or stagna). The hydrological shift in La Molina may have been amplified by wet events such as those reconstructed in PVO for the Roman period (Schellekens et al., 2011).

Some taxa such as Corylus and Erica t., as well as the grazing index also show a marked seesaw pattern during Early Roman Empire chronology (Pz-2a) (Fig. 6). Phases with increases in atmospheric metal pollution, pointing to changes in the intensity of mining/metallurgy, are coeval with decreases in tree pollen percentages indicating episodes of woodland clearance (Fig. 6). Those episodes, which coincided with increased atmospheric metal pollution and grazing pressure, were mostly connected with decreases in Corylus and Erica t (Fig. 7). Pliny the Elder, who served as procurator in Hispania, described the way Romans used the wood for gold mining (outlined in Reher et al., 2012). Timber was used for shoring in galleries or pits, a technique called aurum canalicium. This technique was infrequently used in the region, although some beams have been found in mining contexts close to La Molina mire, such as the Boinós site (Villa Valdós, 1998). Pliny also documented the use of heather, which was employed to retain the gold in the washing channels (agogae) and burned afterwards in order to extract the gold particles caught in the branches. Heather (Erica t.) values decrease synchronously with the peaks in metal pollution, most likely indicating its use for gold extraction. In fact, the relationships between Corylus and Erica t. with metal pollution show large Pearson's correlation coefficients for the Early Roman Empire

Late Iron Age Early Roman Empire Late Roman Empire Germanic Period

0 0 0 n 0 +-

-0.8 -0.6 -0.4 -0.7 -0.5 -0.3 -0.1 0.1 0.3 0.5 -0.8 -0.6 -0.4 -0.2 0 -0.15 -0.1 -0.05 0

• Corylus x deciduous Quercus o Erica t.

Fig. 7. Correlation between indicators of metal pollution and grazing with deciduous Quercus, Corylus and Erica t. for the Late Iron Age, Early and Late Roman Empire and Germanic Period. Note that metal pollution values haven been multiplied by -1, so the larger the score of a sample, the higher the metal pollution is. This has been done to make the interpretation of the results more intuitive. Pearson's correlation coefficients for the same indicators are in Table 1.

(Table 1); the larger the metal pollution the larger the decreases in hazel and heather were. Similarly, the relationship between hazel and heather with the grazing index shows strong correlations during the Early Roman Empire (Table 1, Fig. 7), because more intense phases of grazing pressure are coincident with the largest decreases in hazel and heather. These relationships are not detected either before or after this period (Fig. 7, Table 1).

The two phases of decline in deciduous Quercus pollen percentages detected during the Early Roman Empire (Pz-2a) are more difficult to interpret, although they are probably related to the disturbance of the woodland cover with repeated clearances due to the presence of local human activities. They could have been related to crops sown very close to the mire, because they coincide with Cerealia t. peaks (Fig. 6). However, the Cerealia t. signal could also be linked to the higher water-table of the mire, as some wild grasses produce pollen of cereal size and grow in aquatic environments (e.g., Glyceria; Joly et al., 2007). Although the fact that deciduous oaks occupy deeper soils suitable for cultivation cannot be neglected, a combination of both processes could explain the large percentages of Cerealia t. Additionally, it is likely that the increase in Castanea percentages and the appearance of Juglans indicate the cultivation of chestnut and walnut (Fig. 4) at a regional scale. Unfortunately, there is no record of Roman settlements in the vicinity (Rodríguez Otero, 1992; Fernaóndez Mier, 1999), whose inhabitants would have farmed locally, except for an inscription found beside the mines (CIL II, 5739). Nonetheless, not far from La Molina mire an old branch of the Way of Saint James pilgrimage route (Fig. 2) passes by, which could have been a communication route in ancient times. The intensification of agricultural indicia and the significant presence of Roman activity in the area allow us to expect new sites to be found in the immediate surroundings with further archaeological research.

For the Late Roman Empire (Pz-2b, AD c. 240—440), any differentiation in the use of resources linked to mining activities is less evident, as the decline of hazel and heather are no longer coeval (Fig. 7, Table 1). The mire water-table was much lower (Fig. 3), in

agreement with archaeological evidence that gold mining was only important during the Early Roman Empire (Domergue, 1990). Even so, atmospheric metal pollution was still high and probably related to metallurgy more broadly rather than specifically to gold mining, although other mines far from La Molina could have been exploited. Metalworking, together with farming activities, would explain the woodland decline. In fact, a woodland clearance event affecting both deciduous Quercus and Corylus is coincident with a peak in grazing pressure and cereal cultivation along with indication of metal pollution (Fig. 6). Additionally, Juglans pollen is almost absent, but Castanea has similar values to those recorded during the Early Roman Empire, suggesting that it continued to be cultivated (Fig. 4).

The Germanic period (Pz-3, AD c. 440—700) was also a phase of change in land use. The seesaw pattern, which started to disappear in the Late Roman Empire, is no longer evident either for forest clearances, atmospheric metal pollution, soil erosion or water-table fluctuation (Figs. 3, 6 and 7). Permanent grazing is the most prominent feature while Cerealia t. is almost absent. Forest clearance took place to create pasture rather than for cultivation. Nevertheless, the effect of human activities on the woodland cover was not so strong as to prevent some regeneration, a pattern also observed during the local Late Iron Age and Roman periods. This indicates that human activities, although quite intense in some phases, were probably of local scale. The human impact on woodlands was not strong enough to trigger irreversible changes after the perturbations, principally the intense mining activities detected during the Early Roman Empire. The same conclusion can be extracted from the rarefaction analysis. Palynological richness values increased during the Early Roman Empire, when local human impact was more intense, although from the Late Roman Empire to Germanic times they gradually return to values similar to those found during the Late Iron Age (Figs. 5 and 6). In the study performed on a peat core from the Ayoó de Vidriales mire (Northern Iberia Plateau; Morales-Molino and García Antón, 2014) an increase in the palynological richness with increasing human disturbance was also found.

The archaeobotanical work performed in the Las Medulas Roman mining landscape (Leon province) on pre-Roman, Roman and post-Roman archaeological sites showed the importance of Casta-nea, higher presence of both Cerealia t. pollen and carpological remains of emmer (Triticum dicoccum), barley (Hordeum vulgare), broomcorn millet (Panicum miliaceum) and broad beans (Vicia faba), and lower values of mesophilous trees in Roman times (Lopez-Merino et al., 2010a). After Roman times, cereal crops decreased and local grazing indicators increased, as shown by the lower percentages of Cerealia t. and the higher values of coproph-ilous fungi (Loepez-Merino et al., 2008). Unfortunately, there are not many continuous palynological records on natural archives with good chronological control and high-resolution for these chronologies, close to Roman mining areas, to develop a more general picture of the impact of human activities. One exception is located in the Teleno Mountains close to Las Meedulas: the Xan de Llamas mire at 1500 m asl (Morales-Molino et al., 2011). It shows a deforestation phase with a high incidence of fires from c. 2850 to 1300 cal yr BP, i.e., from the Iron Age to Germanic times. After this phase woodlands recovered. Hence, even in places such as Las Medulas in which the impact of mining was very dramatic at many scales, forests recovered when the extractive activities ceased.

The changes in vegetation described above are consistent with those recorded in pollen diagrams from other later prehistoric and Roman mining and/or metalworking regions in Europe (i.e., Le Roux et al., 2004; Mighall et al., 2010, 2012). At all sites woodland clearance associated with metal pollution and arable and livestock farming commenced in the Late Iron Age and intensified during the Roman period. Miners and/or metallurgists appear to have exploited local woodland, a hypothesis that is also supported by anthracological data (Ludemann et al., 2004; Ludemann, 2010). Any preference for particular tree species appears to have been simply controlled by local availability: the most abundant trees and shrubs were exploited, such as hazel and heather in the case of La Molina. Phases of forest regeneration also typically occur when mining activities cease: at La Molina percentages of Corylus, Erica t., and occasionally Quercus increase in between the phases of metal pollution. As pointed out by other studies in other historical mining areas, the impact of mining on woodlands was short-lived (Monna et al., 2004a, 2004b; Breitenlechner et al.,2010; Mighall et al.,2012; Reher et al., 2012).

5.2. Non-regeneration of forests since the Middle Ages

In the palaeoenvironmental record of La Molina mire the episodes of forest clearances detected from the Late Iron Age to Germanic times were small-scale and short-lived, after which forests recovered. Thus, although mining and metallurgy exerted a large control on the resources and a new regime to manage the landscape, such activities did not trigger major changes at a landscape scale, neither during the intense rise of Roman mining/ metallurgical activities nor after they ceased, because the Visi-gothic lifestyle that developed after the Late Roman Empire was a subsistence farming structure mainly based on livestock and fruit trees (i.e., oak and chestnut) (Valbuena-Carabana et al., 2010). Thus, relatively extensive woodland still persisted until the Middle Ages.

The transition between Germanic times (Pz-3) and the Middle Ages (Pz-4, AD c. 700-1450) led to the onset of a continuous, gradual and permanent deforestation (Fig. 4), which fits with the widespread reduction of the forest cover since Medieval times detected in other high-resolution studies performed in NW Spain: for example, at AD c. 700-800 in the Ancares, Courel and Xistral Mountains (Munoz Sobrino et al., 1997, 2001; Mighall et al., 2006), at AD c. 600-700 in the Monte Areo and Bocelo

Ranges (Lopez-Merino et al., 2010b; Silva Sanchez et al., 2014), and at AD c. 500—1000 in the Monte Paradela (Carrión et al., 2010; Kaal et al., 2011; López-Merino et al., 2012). This loss of woodlands is one of the most dramatic changes in the vegetation history of NW Spain (Martínez Cortizas et al., 2005). Previously forested areas seem to have been occupied by the characteristic heathlands that compose the NW Iberian landscape today (Kaal et al., 2011). The high percentages of Erica t., C. vulgaris, Cyti-sus/Ulex t. and Prunus t. found in the pollen record of La Molina also support this interpretation (Fig. 4). But grasslands, as inferred by the higher percentages of Poaceae, were also very important. Although indicators of local grazing activities (i.e., coprophilous fungi) are not significant in La Molina record during the Middle Ages (Fig. 3), transhumance livestock was notable at a regional scale and is considered to be one of the main drivers for the deforestation process in the whole Spanish territory (Gil Sanchez, 2011). Although no microcharcoal information is available for the La Molina peat record, which would have been interesting to support the regional scale character of the forest decline and the grassland expansion, the study done by Mighall et al. (2006) in the nearby Pena da Cadela bog (PDC, 974 m asl, Xistral Mountains), located c. 100 km west of La Molina, shows an increase in the microcharcoal content since c. 700 AD, which supports the broad-scale of the Middle Ages deforestation process in NW Iberian landscapes.

Consequently, the Middle Ages supposed a large-scale, progressive change in the landscape, as deforestation occurred at a broader scale across the whole Iberian Peninsula for many purposes in a time of instability, characterised by Christian/Muslims wars and invasions, which impeded forest regeneration. The principal causes were related with mobile livestock, mainly sheep, which triggered the burning of woodlands for the creation of extensive pastures, but also shipbuilding and caulking, charcoal extraction, and the intensification of cultivation (Valbuena-Carabana et al., 2010). However, at the mire local scale the changes that happened at the onset of Roman times in La Molina seem to have been crucial for the future dynamics of the mire. Changes caused by higher water-table levels were irreversible, triggering a shift towards a new trophic status (ombrotrophy) (Fig. 3). Thus, responses at regional and local scales were related to different environmental stressors, highlighting the complexity of landscape trajectories.

Since the Middle Ages other human activities were also important in shaping the landscape surrounding La Molina. These include the consolidation of the cultivation of sweet chestnut (Conedera et al., 2004) since Pz-4, as well as the widespread pine and eucalyptus afforestation across NW Iberia (Sande Silva, 2007) and the local equine and bovine grazing during the last two centuries around and on the La Molina mire itself (Pz-5b). Additionally, the La Molina record shows a rise in atmospheric metal pollution, starting in the Early Modern period and increasing dramatically since the onset of the Industrial Revolution (Fig. 3), which is the same as that typically detected for other records in NW Spain (Martínez Cortizas et al., 2002, 2012; Kylander et al., 2005; Olid et al., 2010). This was the first time in more than 2000 years that atmospheric pollution reached values higher than those detected during Roman times. Interestingly, palynological richness presents more or less constant values during the Middle Ages (Pz-4) and the Early Modern period (Pz-5a), while values sharply increase at AD c. 1850 (P-5b, Fig. 5). Palynological richness seems to be more sensitive to local perturbations on the landscape rather than to regional ones: the first increase during the Early Roman Empire is linked to the local mining, while the most recent increase is coeval with the local afforestation with pines and eucalyptus (Fig. 4).

6. Conclusions

The palynological study of La Molina mire has proved to be a powerful tool to identify and reconstruct landscape changes related to human activities since the Late Iron Age in NW Iberia. The mire studied, La Molina, is particularly well suited for this purpose because it is located in an area of the Spanish Asturias province that was intensively mined for gold in Roman times. Gold mining used hydraulic power and the intense fluctuations in the mire's hydrology started at the beginning of the Early Roman Empire (c. 20 BC-AD 240), when Roman domination began in the area, pointing to its use as a water reservoir in new extraction works. Synchronous clearances of the local woodlands to extract wood resources to develop mining/metalworking and to open spaces for cultivation and livestock grazing were also detected. The most striking feature is that the use of natural resources seems to have been much more complex and variable during Roman than in pre-Roman (local Late Iron Age) and post-Roman (Germanic Period) times. This fact suggests different strategies of landscape management and socio-environmental interactions. In fact, the exceptionally detailed record for the Early Roman Empire period enabled the identification of these contrasting strategies, e.g., different forest clearances for several purposes. The use of hazel and heather seems to have been linked for mining/metalworking during the Early Roman Empire.

Although the impact of mining and other human activities from the late Iron Age to Germanic times, and mostly during the Early Roman Empire, was intense and affected the landscape, these early effects were seemingly reversible, because the forest regenerated after mining ceased. These forest clearances were local and shortlived. In contrast, the onset of the Middle Ages involved a nonreturn change, because deforestation was generalised across the landscape as a whole and linked to many different human activities.

Acknowledgements

This work was funded by the projects HAR2008-06477-C03-03/ HIST, CGL2010-20672 (Plan Nacional I+D+i, Spanish Ministry of Science and Innovation), 10PXIB200182PR (General Directorate of I+D, Xunta de Galicia), and CDS-TCP (CSD2007-00058, Programa Consolider-Ingenio 2010). We are grateful to Fernando Gil Sendino, Carmen Fernandez Ochoa and Roberto Zapico for their financial support, collaboration and assistance during the field work, to Margarita Fernandez Mier, Angel Villa Valdas and Jorge Camino Mayor for the information about the study area and to Suzanne Leroy for her comments and advice on earlier versions of the manuscript. Thanks also to Dr Caesar Morales-Molino and two anonymous referees for their detailed comments on an earlier version of the manuscript.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.jas.2014.07.016.

References

Birks, H.J.B., 2012. Ecological palaeoecology and conservation biology: controversies, challenges, and compromises. Int. J. Biodivers. Sci. Ecosyst. Serv. Manag. 8, 292—304.

Blaauw, M., 2010. Methods and code for 'classical' age-modelling of radiocarbon

sequences. Quat. Geochronol. 5, 512—518. Breitenlechner, E., Hilber, M., Lutz, J., Kathrein, Y., Unterkircher, A., Oeggl, K., 2010. The impact of mining activities on the environment reflected by pollen, charcoal and geochemical analyses. J. Archaeol. Sci. 37,1458—1467. Camino Mayor, J., 1989a. Los Remoleiros. Carta Arqueológica de Asturias. Principado de Asturias.

Camino Mayor, J., 1989b. Las Mueyes. Carta Arqueológica de Asturias. Principado de Asturias.

Carrión, J.S., Navarro, C., Navarro, J., Munuera, M., 2000a. The interpretation of cluster pine (Pinus pinaster) in floristic-phytosociological classifications from a palaeoecological perspectiva. Holocene 10, 243—252.

Carrión, J.S., Parra, I., Navarro, C., Munuera, M., 2000b. The past distribution and ecology of the cork oak (Quercus suber) in the Iberian Peninsula: a pollen-analytical approach. Divers. Distrib. 6, 29—44.

Carrión, Y., Kaal, J., Lopez-Saez, J.A., López-Merino, L., Martínez Cortizas, A., 2010. Holocene vegetation change in NW Spain revealed by anthracological and palynological records from a colluvial soil. Holocene 20, 53—66.

Conedera, M., Krebs, P., Tinner, W., Pradella, M., Torriani, D., 2004. The cultivation of Castanea sativa (Mill.) in Europe, from its origin to its diffusion on a continental scale. Veg. Hist. Archaeobot. 13,161—179.

Curras, A., Zamora, L., Reed, J.M., García-Soto, E., Ferrero, S., Armengol, X., Mezquita-Joanes, F., Marqués, M.A., Riera, S., Julia, R., 2012. Climate change and human impact in central Spain during Roman times: high-resolution multi-proxy analysis of a tufa lake record (Somolinos, 1280 m asl). Catena 89, 31—53.

Dearing, J.A., 2006. Climate-human-environment interactions: resolving our past. Clim. Past 2,187—203.

Dearing, J.A., Battarbee, R.W., Dikau, R., Larocque, I., Oldfield, F., 2006. Humanenvironment interactions: learning from the past. Reg. Environ. Change 6,1—16.

Domergue, C., 1987. Catalogue des mines et des fonderies antiques de la Póninsule Iberique. Casa de Velózquez-Diffusion de Boccard, Madrid.

Domergue, C., 1990. Les mines de la póninsule Ibórique dans l'antiquitó romaine. Ecole française de Rome, Rome.

Domergue, C., Herail, G., 1978. Utilisation des vestiges archóologiques dans la reconstitution de l'eóvolution des milieux. L'exemple des mines romaines du Nord-Ouest de l'Espagne. Caesarodunum 13, 227—239.

Edmondson, J.C., 1989. Mining in the later roman Empire and beyond: continuity or disruption? J. Roman Stud. 79, 84—102.

Fanjul Peraza, A., Menóndez Bueyes, L.R., 2003—2007. "Antiguas" y canales. El complejo minero romano de Les Mueches-Ablaneda (Salas, Asturias). Niv. Cero 11, 79—94.

Fernaóndez Mier, M., 1999. Geónesis del territorio en la Edad Media. Arqueología del paisaje y evolución histórica de la montana asturiana: el valle del río Pigüena. Universidad de Oviedo, Oviedo.

Gil Sónchez, L., 2011. El bosque que nos ha llegado: la extinción local de los bosques en Espana. In: Ezquerra Boticario, F.J., Rey van den Bercken, E. (Eds.), La evolucioón del paisaje vegetal y el uso del fuego en la Cordillera Cantóabrica. Fundación Patrimonio Natural de Castilla y León, León, pp. 153—170.

Hammer, 0., Harper, D.A.T., Ryan, P.D., 2001. PAST: paleontological statistics software package for education and data analysis. Palaeontol. Electron. 4, 9. http:// palaeo-electronica.org/2001_1/past/issue1_01.htm.

Hong, S., Candelone, J.-P., Patterson, C.C., Boutron, C.F., 1994. Greenland ice evidence of hemispheric lead pollution two millennia ago by Greek and Roman civilizations. Science 265,1841—1843.

Joly, C., Barillóe, L., Barreau, M., Mancheron, A., Visset, L., 2007. Grain and annulus diameter as criteria for distinguishing pollen grains of cereals from wild grasses. Rev. Palaeobot. Palynol. 146, 221—233.

Jouffroy-Bapicot, I., Pulido, M., Baron, S., Galop, D., Monna, F., Lavoie, M., Ploquin, A., Petit, C., de Beaulieu, J.-L., Richard, H., 2007. Environmental impact of early palaeometallurgy: pollen and geochemical analysis. Veg. Hist. Archaeobot. 16, 251—258.

Kaal, J., Carrioón Marco, Y., Asouti, E., Martín Seijo, M., Martínez Cortizas, A., Costa Casóis, M., Criado Boado, F., 2011. Long-term deforestation in NW Spain: linking the Holocene fire history to vegetation change and human activities. Quat. Sci. Rev. 20, 161—175.

Kylander, M.E., Weiss, D.J., Martínez Cortizas, A., Spiro, B., Garcia-Sanchez, R., Coles, B.J., 2005. Refining the pre-industrial atmospheric Pb isotope evolution curve in Europe using an 8000 year old peat core from NW Spain. Earth Planet. Sci. Lett. 240, 467—485.

Le Roux, G., Weiss, D., Grattan, J., Givelet, N., Krachler, M., Cheburkin, A., Rausch, N., Kober, B., Shotyk, W., 2004. Identifying the sources and timing of ancient and medieval atmospheric lead pollution in England using a peat profile from Lindow bog, Manchester. J. Environ. Monit. 6, 502—510.

Loópez-Merino, L., 2009. Paleoambiente y antropizacioón en Asturias durante el Holoceno. Universidad Autoónoma de Madrid (PhD thesis).

López-Merino, L., López-Saez, J.A., Abel Schaad, D., Sónchez-Palencia, F.J., Reher Díez, G.S., 2008. Dinaómica antroópica en El Bierzo (Leóon) desde óepoca romana: estudio palinoloógico de Castro Ventosa. Polen 18, 25—36.

López-Merino, L., Pena-Chocarro, L., Ruiz-Alonso, M., Lopez-Saez, J.A., Sanchez-Palencia, F.J., 2010a. Beyond nature: the management of a productive cultural landscape in Las Meódulas area (El Bierzo, Leoón, Spain) during pre-Roman and Roman times. Plant Biosyst. 144, 905—919.

López-Merino, L., Martínez Cortizas, A., López-Saez, J.A., 2010b. Early agriculture and palaeoenvironmental history in the North of the Iberian Peninsula: a multi-proxy analysis of the Monte Areo mire (Asturias, Spain). J. Archaeol. Sci. 37,1978—1988.

López-Merino, L., Martínez Cortizas, A., López-Saez, J.A., 2011. Human-induced changes on wetlands: a study case from NW Iberia. Quat. Sci. Rev. 30, 2745—2754.

López-Merino, L., Silva Sónchez, N., Kaal, J., López-Sóez, J.A., Martínez Cortizas, A., 2012. Post-disturbance vegetation dynamics during the Late Pleistocene and the Holocene: an example from NW Iberia. Glob. Planet. Change 92—93, 58—70.

Ludemann, T., 2010. Past fuel wood exploitation and natural forest vegetation in the Black Forest, the Vosges and neighbouring regions in western Central Europe. Palaeogeogr. Palaeoclimatol. Palaeoecol. 291,154—165.

Ludemann, T., Michiels, H.-G., Noken, W., 2004. Spatial patterns of past wood exploitation, natural wood supply and growth conditions: indications of natural tree species distribution by anthracological studies of charcoal-burning remains. Eur. J. For. Res. 123, 283—292.

Marín Suórez, C., 2011. La Edad del Hierro en el Occidente Cantóbrico: de la cultura arqueológica al grupo arqueológico. Fervedes 7,123—132.

Martínez Alcíbar, A., 1850. Examen de antiguos trabajos de explotación de minerales auríferos en Asturias y noticias sobre la Ballesterosita y la Plumbostannita. Rev. Minera 1, 33—50.

Martínez Cortizas, A., Potevedra-Pombal, X., Nóvoa-Munoz, J.C., García-Rodeja, E., 1997. Four thousand years of atmospheric Pb, Cd and Zn deposition recorded by the ombrotrophic peat bog of Penido Vello. Water Air Soil Pollut. 100, 387—403.

Martínez Cortizas, A., Potevedra-Pombal, X., García-Rodeja, E., Nóvoa-Munoz, J.C., Shotyk, W., 1999. Mercury in a Spanish peat bog: archive of climate change and atmospheric metal deposition. Science 284, 939—942.

Martínez Cortizas, A., García-Rodeja, E., Pontevedra Pombal, X., Nóvoa-Munoz, J.C., Weiss, D., Cherbulin, A., 2002. Atmospheric Pb deposition in Spain during the last 4600 years recorded by two ombrotrophic peat bogs and implications for the use of peat as archive. Sci. Total Environ. 292, 33—44.

Martínez Cortizas, A., Mighall, T., Pontevedra-Pombal, X., Nóvoa Munoz, J.C., Pei-teado Varela, E., Pineiro Rebolo, R., 2005. Linking changes in atmospheric dust deposition, vegetation change and human activities in northwest Spain during the last 5300 years. Holocene 15, 698—706.

Martínez Cortizas, A., Peiteado Varela, E., Bindler, R., Biester, H., Cheburkin, A., 2012. Reconstructing historical Pb and Hg pollution in NW Spain using multiple cores from the Chao de Lamoso bog (Xistral Mountains). Geochim. Cosmochim. Acta 82, 68—78.

Martínez Cortizas, A., López-Merino, L., Bindler, R., Mighall, T., Kylander, M., 2013. Atmospheric Pb pollution in N Iberia during the late Iron Age/Roman times reconstructed using the high-resolution record of La Molina mire (Asturias, Spain). J. Paleolimnol. 50, 71 —86.

Mighall, T.M., Chambers, F.M., 1997. Early ironworking and its impact on the environment: palaeoecological evidence from Bryn y Castell hillfort, Snowdonia, North Wales. Proc. Prehist. Soc. 63, 199—219.

Mighall, T.M., Martínez Cortizas, A., Biester, H., Turner, S.E., 2006. Proxy climate and vegetation changes during the last five millennia in NW Iberia: pollen and nonpollen palynomorph data from two ombrotrophic peat bogs in the North Western Iberian Peninsula. Rev. Palaeobot. Palynol. 141, 203—223.

Mighall, T.M., Timberlake, S., Crew, P., 2010. Vegetation changes in former mining and metalworking areas of Wales and Ireland during prehistoric and medieval times. In: Belford, P., Palmer, M., White, R. (Eds.), Footprints of Industry, British Archaeological Reports Series, 523, pp. 19—26.

Mighall, T.M., Chambers, F.M., Timberlake, S., O'Brien, W.F., 2012. Characterising vegetation changes in former mining and metalworking areas during prehistoric and Roman times. Not. Archeol. Bergomensi 20, 117—130.

Monna, F., Galop, D., Carozza, L., Tual, M., Beyrie, A., Marembert, F., Chateau, C., Dominik, J., Grousset, F., 2004a. Environmental impact of early Basque mining and smelting recorded in a high ash minerogenic peat deposit. Sci. Total Environ. 327,197—214.

Monna, F., Petit, C., Guillaumet, J.-P., Jouffroy-Bapicot, I., Blanchot, C., Dominik, J., Losno, R., Richard, H., Levêque, J., Chateau, C., 2004b. History and environmental impact of mining activity in Celtic Aeduan territory recorded in a peat-bog (Morvan-France). Environ. Sci. Technol. 38, 665—673.

Moore, P.D., Webb, J.A., Collinson, M.E., 1991. Pollen Analysis, second ed. Blackwell Scientific Publications, London.

Morales-Molino, C., García Antón, M., 2014. Vegetation and fire history since the last glacial maximum in an inland area of the western Mediterranean Basin (Northern Iberian Plateau, NW Spain). Quat. Res. 81, 63—77.

Morales-Molino, C., García Anton, M., Morla, C., 2011. Late Holocene vegetation dynamics on an Atlantic-Mediterranean mountain in NW Iberia. Palaeogeogr. Palaeoclimatol. Palaeoecol. 302, 323—337.

Munoz Sobrino, C., Ramil-Rego, P., Rodríguez Guitián, M., 1997. Upland vegetation in the north-west Iberian Peninsula after the last glaciation: forest history and deforestation dynamics. Veg. Hist. Archaeobot. 6, 215—233.

Munoz Sobrino, C., Ramil-Rego, P., Rodríguez Guitian, M., 2001. Vegetation in the mountains of northwest Iberia during the last glacial-interglacial transition. Veg. Hist. Archaeobot. 10, 7—21.

Olid, C., García-Orellana, J., Martínez Cortizas, A., Masqueó, P., Peiteado Varela, E., Sónchez-Cabeza, J.A., 2010. Multiple site study of recent atmospheric metal (Pb, Zn and Cu) deposition in the nW Iberian Peninsula using peat cores. Sci. Total Environ. 408, 5540—5549.

Orejas, A., 1994. Le main d'oeuvre des mines d'or romaines du Nord-Ouest de la Póninsule Ibórique. Centre de Recherches d'Histoire Ancienne. Faculte de Lettres. Besançon, Universitó de Franche-Comtó.

Orejas, A., Sóanchez-Palencia, F.J., 2002. Mines, territorial organization, and social structure in Roman Iberia. Am. J. Archaeol. 106, 581—599.

Paillette, A., 1852. Recherches sur l'histoire et les conditions de gisement des mines d'or dans le nord de l'Espagne. Bull. Soc. góol. Fr. 2ème sór. 9, 482—504.

Perea, A., Sanchez-Palencia, F.J., 1995. Arqueología del oro Astur: orfebrería y minería. Caja de Asturias, Obra Social y Cultural, Oviedo.

Pontevedra-Pombal, X., Mighall, T.M., Nóvoa-Munoz, J.C., Peiteado-Varela, E., Rodríguez-Racedo, J., García-Rodeja, E., Martínez-Cortizas, A., 2013. Five thousand years of atmospheric Ni, Zn, As, and Cd deposition recorded in bogs from NW Iberia: prehistoric and historic anthropogenic contributions. J. Archaeol. Sci. 40, 764—777.

Reher, G.S., Loópez-Merino, L., Sóanchez-Palencia, F.J., Loópez-Sóaez, J.A., 2012. Configuring the landscape: Roman mining in the conventus Asturum (NW Hispania). In: Kluiving, S., Guttmann-Bond, E. (Eds.), Landscape Archaeology between Art and Science: From a Multi- to an Interdisciplinary Approach. Amsterdam University Press (Landscape & Heritage Proceedings), Amsterdam, pp. 125—134.

Reille, M., 1992. Pollen et Spores d'Europe et d'Afrique du Nord. Laboratoire de Botanique Historique et Palynologie, Marseille.

Rodríguez Otero, V., 1992. Carta arqueológica del concejo de Salas. Enero — Octubre 1989. In: Excavaciones arqueológicas en Asturias 1987—90. Principado de Asturias, Oviedo.

Sónchez-Palencia, F.J., 1983. La explotación del oro de Asturia y Gallaecia en la antigüedad. Universidad Complutense de Madrid (PhD thesis).

Sónchez-Palencia, F.J., 1995. Minería y metalurgia de la región astur en la antigüedad. In: Fernandez Ochoa, C. (Ed.), Astures. Pueblos y culturas en la frontera del imperio romano. Caja de Asturias, Oviedo.

Sóanchez-Palencia, F.J. (Ed.), 2000. Las Meódulas (Leoón). Un paisaje cultural en la Asturia Augustana. Instituto Leonós de Cultura (Diputación Provincial de León), León.

Sónchez-Palencia, F.J., Suórez, V., 1985. La minería antigua del oro en Asturias. In: Hunosa (Ed.), Libro de la mina. Asturias, pp. 165—169.

Sónchez-Palencia, F.J., López-Saez, J.A., Reher Díaz, G.S., López-Merino, L., 2011. La minería romana en Leon y Asturias, su importancia en la configuration de los paisajes. In: Ezquerra Boticario, F.J., Rey van den Bercken, E. (Eds.), La evolución del paisaje vegetal y el uso del fuego en la Cordillera Cantaóbrica. Fundacioón Patrimonio Natural de Castilla y León, León, pp. 125—138.

Sande Silva, J., 2007. Pinhais e eucaliptais. A foresta cultivada. Público, Comunicaçâo, S.A. Fundacâo Luso-Americana para o desenvolvimento, Lisboa.

Schellekens, J., Buurman, P., Fraga, I., Martínez Cortizas, A., 2011. Holocene vegetation and hydrologic changes inferred from molecular vegetation markers in peat, Penido Vello (Galicia, Spain). Palaeogeogr. Palaeoclimatol. Palaeoecol. 299, 56—69.

Schulz, G., Paillette, A., 1850. Notice sur une pyrite stannifère (Ballestórosite) et sur quelques gisements d'ótain en Espagne. Bull. la Soc. góol. Fr. 2ème sór. 7,16—25.

Silva Sónchez, N., Martínez Cortizas, A., López-Merino, L., 2014. Linking forest cover, soil erosion and mire hydrology to Late Holocene human activity and climate in NW Spain. Holocene 24, 714—725.

Valbuena-Carabana, M., López de Heredia, U., Fuentes-Utrilla, P., Gonzalez-Doncel, I., Gil, L., 2010. Historical and recent changes in the Spanish forests: a socio-economic process. Rev. Palaeobot. Palynol. 162, 492—506.

Villa Valdós, A., 1998. Estudio arqueológico del complejo minero romano de Boinas, Belmonte de Miranda (Asturias). Bol. Geol. Y Min. 109, 589—598.