Holocene palaeoecology and floodplain evolution of the Muge tributary, Lower Tagus Basin, Portugal

ARTICLE IN PRESS

Quaternary International 189 (2008) 135–151

Holocene palaeoecology and floodplain evolution of the Muge tributary, Lower Tagus Basin, Portugal Tim van der Schrieka,, David G. Passmorea, Fatima Franco Mugicab, Anthony C. Stevensona, Ian Boomera,1, Jose Rola˜oc a

School of Geography, Politics and Sociology, Daysh Building, University of Newcastle upon Tyne, Newcastle upon Tyne NE1 7RU, UK b Department of Ecology, Universidad Auto´noma de Madrid, Campus de Cantoblanco, 28049 Madrid, Spain c Department of Archaeology, Universidade Auto´noma de Lisboa, Pala´cio dos Condos de Redondo, Rua Santa Marta 47, 1169-023 Lisbon, Portugal Available online 7 September 2007

Abstract The lower Muge valley, a tributary of the lower Tagus River (Portugal), features an important archaeological record of Mesolithic shell midden sites. Archaeological research has long assumed that tidally influenced valley floor environments in the immediate locality of the sites provided a rich food resource, attracting Mesolithic settlement. To date there has been little attempt to use palaeoenvironmental records to reconstruct Holocene floodplain evolution in the Lower Tagus valley. The cultivated freshwater lower Muge floodplain is locally underlain by 11 m of homogeneous fine-grained sediments and peat, comprising buried floodplain environments contemporary with Mesolithic occupation (6200–4800 cal BC). Pollen and foraminifera analyses demonstrate that fine-grained deposition, forced by sea-level rise, commenced 6200 cal BC in an estuarine setting. The lower Muge floodplain experienced maximum tidal influence 5800–5500 cal BC. Subsequently, sediment supply rates overtook the decreasing rate of sea-level rise and fluvial environments expanded. The pollen record may suggest regional desiccation from 5000 cal BC. Estuarine environments disappeared suddenly 3800 cal BC when freshwater wetlands were established. Although the initiation of Mesolithic settlement is shown to coincide with the beginning of tidal influence, site abandonment does not match with any major environmental change. Sea-level still stand (2600 cal BC) has been linked to valley floor stabilisation and soil formation. Alder floodplain woodland developed prior to 230 cal BC and was cleared, probably during Roman times, for agriculture. Renewed deposition after 230 cal BC may relate to internal mechanisms or to human impact upon the catchment vegetation. r 2007 Elsevier Ltd and INQUA. All rights reserved.

1. Introduction The lower Tagus valley has a rich, well-preserved archaeological record of mid-Mesolithic shell middens that are internationally significant in European prehistory (Straus et al., 1990; Zilha˜o, 1993, 2000; Vierra, 1995). Particular importance is attributed to the main cluster of five Mesolithic shell middens along the lower Muge tributary (Fig. 1) which has been studied over the past 140 years (e.g. Mendes Correa, 1933; Roche, 1977; Morais Arnaud, 1989; Rola˜o, 1994; Cunha and Cardoso, 2003). Corresponding author. Tel.: +44 (0) 191 222 6757.

E-mail address: [email protected] (T. van der Schriek). Present address: School of Geography, Earth and Environmental Sciences, The University of Birmingham, Edgbaston, Birmingham B15 2TT, UK. 1

The surviving midden sites are up to 5 m high and 100 m in diameter, and were probably occupied 6200–4800 cal BC (Zilha˜o, 2000). Faunal and isotope studies testify to the mixed marine–terrestrial resources exploited by the Mesolithic communities, with marine resources accounting for 50% of the diet (e.g. Lentacker, 1986; Lubell et al., 1994; Richards and Hedges, 1999). Although Mesolithic settlement subsistence and the local occupation history is commonly linked to the establishment and subsequent cessation of inner estuarine conditions in the adjacent valley floor (e.g. Straus et al., 1990; Zilha˜o, 1993, 2000; Bicho, 1994), published interpretations of the environmental context of the middens have been solely based on faunal analyses at site level (e.g. Veiga Ferreira, 1954; Lentacker, 1986; Detry, 2001) and on palynological records of adjacent regions (e.g. Morais Arnaud, 1989; Bicho, 1994; Vierra, 1995). There has been little attempt to use

1040-6182/$ - see front matter r 2007 Elsevier Ltd and INQUA. All rights reserved. doi:10.1016/j.quaint.2007.09.007

ARTICLE IN PRESS T. van der Schriek et al. / Quaternary International 189 (2008) 135–151

136

0

Tagus River

2km

1

1 2 3

5 11

12 13

20

10

40

4

48 Raposa

64 5

51

Muge

Auger core

77

Spain

Dissected Tertiary Basin Fill (101m<)

Cobra core Terrace Levels (7-95m)

Machine excavated trench

Location of transect (Fig. 2) Shell midden (1: Fonte do Padre Pedro, 2: Florda Beira,3: Cabeço da Arruda, 4: Moita do Sebastiao,5: Cabeçoda Amoreira)

Portugal

R.Tagus

Coalescent Alluvial fans

Bank Exposure

Valley Floor

Muge catchment Lisbon

Active alluvial fans Village Muge

Tertiary Tagus Basin

0

200 km

Fig. 1. Schematic geomorphological map of the lower Muge valley including location of Mesolithic shell midden sites and sediment observations. Labelled cores are discussed in the text. Altitudes are given in metres above Mare´grafo de Cascais (Portuguese datum level).

geoarchaeological methods to reconstruct Holocene physical landscapes and their associated vegetation assemblages that would have been inhabited and exploited by local Mesolithic and later communities. Furthermore, conventional archaeological field surveys are liable to be ineffective across alluvial valley floors where archaeological sites of Mesolithic and later date may have been eroded, reworked and/or deeply buried beneath alluvial sediments. To date, there are no established chronostratigraphic frameworks for late Quaternary valley fill deposits in the lower Tagus valley. We present the latest results of investigations into the changing environmental context of the Mesolithic sites in the lower Muge valley, Portugal (Fig. 1). This paper aims to (i) develop a preliminary chronostratigraphic framework for the lower Muge valley fill and palaeoecological record, and (ii) reconstruct tidal influence and factors controlling Holocene fine-grained infill of the lower Muge valley. 2. Study area The research has focused on the lower reaches of the River Muge, an east-bank tributary of the lower Tagus River draining the central part of the Tertiary Lower Tagus Basin and located 60 km upstream of Lisbon. The Muge River is 55 km long and drains a catchment area of 616 km2; its confluence with the Tagus River is 28 km upstream of the present tidal limit in the lower Tagus valley. The valley of the Muge River is incised into Quaternary alluvial terrace deposits of the Tagus River,

and into Tertiary deposits (Fig. 1). The E–W course of the lower Muge is probably determined by a minor fault perpendicular to the Tagus valley (Barbosa, 1995). The course of the lower Tagus River is determined by a series of NNW–SSE orientated faults that are part of a half-graben structure believed to have been uplifting since the late Tertiary (Barbosa, 1995; Cabral, 1995). The valley floor is bordered to the west by limestone hills up to 600 m above sea level, while Quaternary alluvial deposits border the valley floor to the east in a zone up to 15 km wide (Zbyszewski, 1946; Mozzi et al., 2000). Further to the east, and underlying the depositional terrace levels are unconsolidated clastic deposits of the Tertiary Lower Tagus basin fill, which have experienced between 150 and 200 m of uplift over the past 2–3 Ma (Barbosa, 1995; Cabral, 1995). 3. Methods Geomorphological mapping of the lower Muge area was achieved using aerial photographs, geological and topographical maps and field survey. Lithostratigraphic field description of sediments infilling the lower Muge valley was conducted on deposits exposed by machine excavation (up to 3 m deep) or sediment coring using Eijkelkamp augers of 3–6 cm diameter and 50–100 cm length. Investigations focused on the 10 km long lowland alluvial reach of the Muge River, extending from the Tagus confluence upstream to the village of Raposa; 65 sediment cores, 5 bank exposures and 19 machine excavated trenches were analysed over this reach (Fig. 1).

ARTICLE IN PRESS T. van der Schriek et al. / Quaternary International 189 (2008) 135–151

Selected sediment cores were retaken with a Cobra/Stitz percussion corer, allowing continuous recovery of sediment, and returned to the laboratory for sampling and detailed lithostratigraphic analyses. Selected samples were analysed, using standard techniques, for molluscs (Barret and Younge, 1958; Tucker Abbott, 1990; Peacock, 1993), diatoms (Battarbee, 1986, pp. 528–531) and foraminifera (Brasier, 1980, pp. 162–168). Foraminifera were examined under a reflected light microscope with a 100  magnification. Pollen was extracted from the sediment using conventional pollen preparation techniques including acetolysis and treatment with hydrofluoric acid (Moore et al., 1989); pollen types were identified at 400  magnification, and at 1000  for critical identifications. Total pollen sum was always above 300 grains. Organic content was determined by loss-on-ignition (Heiri et al., 2001) and grain size samples were prepared using standard techniques (Pape, 1996; Konert and Vandenberghe, 1997) prior to analysis by laser with a Malvern Mastersizer S. Lithofacies codes used in this study are based on Miall (1996) and Graham (1988); we developed a lithofacies classification for this study that is based on lithology, fossil content and sedimentary structures evident in cored sediments (Table 1). Dating control is provided by nine 14C dates on in-situ organic materials from representative machine excavated trenches and sediment cores (Table 2). Two of these radiocarbon age estimates dated bulk samples, obtained during reconnaissance survey (Passmore and Stevenson, 1999). For this study, an additional seven samples were extracted at diagnostic levels from the centre of Cobra/Stitz core sections for AMS radiocarbon dating. These samples were deflocculated with analytical grade sodium hexametaphosphate and wet sieved (125 mm mesh) with deionised water; recognisable plant fragments were collected for dating. Throughout this paper, analysed dates are expressed as mid-points of calibrated calendar ages (cal BC/ AD) with age spans at the 2s range. Calibrated ages were calculated with the OxCal 3.5 program (Bronk Ramsey, 1995) using the terrestrial calibration data set (INTCAL98; Stuiver et al., 1998a). One sample (MUG-4), consisting of estuarine mollusc fragments (species Scrobicularia plana), has been calibrated against the marine calibration data set (Marine98; Stuiver et al., 1998b) with DR ¼ 253729 (Monges Soares, 1993). 4. Lower Muge valley fill The presently canalised and cultivated lowland alluvial floodplain of the lower Muge River extends 10 km upstream from the Tagus confluence (Fig. 1). The altitude of its low-relief floodplain varies irregularly from 3 to 5 m above Mare´grafo de Cascais (Portuguese datum level, hereafter named MdC), rendering it below the level of the present Tagus floodplain near the confluence (5–8 m MdC). The width of the tributary floodplain narrows abruptly from 0.5–1.4 km to 0.15 km, 10 km upstream of the

137

Table 1 Descriptive lithofacies codes based on Miall (1996) and Graham (1988) Code

Lithofacies

Sedimentary structures

Sb-ii

Bedding

Sb-i Sb-ie Fl

Sand, medium to coarse, may be pebbly Sand, very fine to coarse Sand, very fine to coarse Sand, silt, mud

Fle

Sand, silt, mud

Fo

Carbonaceous silt, mud

Foe

Carbonaceous silt, mud

Fom

Silt, carbonaceous mud

Fx

Oxidised silt, mud

Fp

Peaty silt, mud

Fpx

Oxidised peaty silt, mud

C Cs

Peat Silty peat

Csx

Oxidised silty peat

A

Sand, silt, mud

Bedding Bedding with estuarine molluscs Laminated, very small scale bedding Laminated, very small scale bedding with estuarine molluscs Finely laminated to massive, with organic inclusions Finely laminated to massive, with organic inclusions and estuarine molluscs Finely laminated to massive, with organic inclusions and frequent mottles Finely laminated to massive, with occasional organic inclusions Finely laminated, frequent plant fragments, mud films Finely laminated to massive, frequent organic inclusions, mud films, blocky structure Plants, plant fragments Plants, plant fragments, finely laminated, mud films Frequent organic inclusions, finely laminated to massive, mud films, blocky structure Antropogenic (made ground)

Tagus confluence, while the floodplain surface rises to above 9 m MdC. This upstream reach of the Muge River displays a valley floor gradient of 2.56 m/km. The lower Muge valley floor is underlain by a wedge of fine-grained clastic sediments and peat that are overly impenetrable coarse sand and gravel deposits. The finegrained valley fill decreases in thickness from 411 m near the Tagus confluence to 3 m at the upstream limit of the lower reach. The fine-grained record lacks distinct erosive boundaries. Present groundwater levels are at 2 m MdC, leaving the greater part of the sedimentary record permanently waterlogged. Nine lithostratigraphic units have been recognised in the fine-grained lower Muge valley fill (Fig. 2). Basal unit 1 (lithofacies Sb-ii, Sb-i) comprises well-sorted bedded coarse sand and gravel with little organic material. The lower boundary of unit 1 remains unknown; only the upper part of this unit has been cored to a maximum depth of 30 cm. The upper contact of unit 1 dips towards the Tagus confluence. Sub-unit 1a has been distinguished on the basis of its geometry from unit 1; its lithology is identical. The sub-unit forms narrow, tabular sand and gravel bodies in a matrix of fine-grained sediment. The lower boundary of unit 1a has not been found; its top was accessed to a maximum depth of 0.5 m. The geometry and

ARTICLE IN PRESS T. van der Schriek et al. / Quaternary International 189 (2008) 135–151

138

Table 2 Radiocarbon dated samples from the fine-grained lower Muge valley fill (Portugal) Sample

Lab. code

14 C age (uncal. BP)

2s Cal. age ranges

Mid-point Cal. age range

Coring site

Sample depth (m)

Sample type

14

C Dating method

Dating event

MUG-1

Beta 111010

2220780

2307180

20

0.94-0.96

Beta 111011

74907180

63257425

20

3.57–3.60

Last period of soil formation Range-finder

MUG-3

AA-49816

7668749

59107120

11

10.73–10.76

Bulk sample peaty silt Bulk sample clayey silt Selected shell fragments

Radiometric

MUG-2

Cal BC 41050 Cal BC 67505900 Cal BC 60305790

AMS

MUG-4

Start of saltwater influence and onset of finegrained AA-48977

7263746

Cal BC 6220–6020

61207100

11

10.64–10.66

Selected plant fragments

AMS

MUG-5

AA-48978

7318744

Cal BC 6240–6060

6150790

20

8.88–8.905

AMS

MUG-6

AA-48979

6626744

Cal BC 5630–5480

5555775

20

4.68–4.70

MUG-7

AA-48980

4985773

Cal BC 3950–3640

37957155

20

200–202.5

MUG-8

AA-48981

5929752

Cal BC 4940–4690

48157125

20

1.365–1.38

MUG-9

AA-48982

3006746

Cal BC 1400–1110, 1100–1080

12407160

40

5.98–6.00

Selected plant and wood fragments Selected plant fragments Bulk sample organic clayey silt Selected wood fragments Selected plant fragments

Radiometric

(Scrobicularia) sedimentation

AMS

AMS

AMS

AMS

Start of saltwater influence (indicated by foraminifera) and onset of fine-grained deposition Start of saltwater influence (indicated by foraminifera) and onset of fine-grained deposition Maximum tidal influence (indicated by foraminiferan peak) End of regional saltwater indicators in pollen diagram Initiation of soil formation and alder invasion (pollen record) Onset of fine-grained sedimentation

Standard preparations and counting procedures of bulk samples MUG-1 and MUG-2 took place at Beta Analytic Inc, Florida. Samples MUG-3–MUG-9 were prepared at the NERC RLC facility and run at the University of Arizona NSF-AMS facility. The graphitisation yield for all AMS samples was at or above the recommended 95% conversion of CO2 to carbon.

E

11

10

10

9

9

8

8 7

7 6

11

5 4

12

20 13

230+/-180 cal BC

3

3795+/-155 cal BC 5555+/-75 cal BC

4m

40

6120+/-100 cal BC

64

5

PZ5

4 3

PZ4 PZ3

PZ2

Unit 1a: Coarse sand-gravel bodies (lithofacies Sb-ii, Sb-i, Fl)

PZ1 20

6150+/-90 cal BC

No Data

Core position (labelled cores are mentioned in text) Base of the fine-grained valley fill

0 0

1km

6

48

m above MdC

m above MdC

W

5910+/-120 cal BC

Lithological Boundary 5005+/-205 cal BC

Radiocarbon Age

Unit 1: Coarse sand and gravel (lithofacies Sb-i, Sb-ii) Unit 2: Laminated-bedded sandy silt (lithofacies Fl, Sb-i) Unit 2a: Laminated-bedded sandy silt with estuarine shells (lithofacies Sb-ie, Fle)

Unit 3: Silty clay (lithofacies Fo, Fom) Unit 3a: Silty clay with estuarine shells (lithofacies Foe) Unit 4: Peat (lithofacies C, Cs) Unit 5: Black peaty clayey silt (lithofacies Csx, Fpx) Unit 6: Silty clay (lithofacies Fo, Fom, Fx)

Fig. 2. Lithostratigraphic model of the lower Muge fine-grained valley fill, based on correlation between the deepest sediment cores in the valley floor; pollen zones are shown for core 20 (see Fig. 8). Accepted radiocarbon dates are included (Table 2). For location of transect and cores see Fig. 1; see Table 1 for lithofacies codes.

ARTICLE IN PRESS T. van der Schriek et al. / Quaternary International 189 (2008) 135–151

8 m near the confluence with the Tagus River. The plastic silty clay of unit 3 contains infrequent sandy laminations and frequent organic material (Fig. 3); its lower boundary with units 2 and 3a is gradual and conformable. Sub-unit 3a (lithofacies Foe) is found in discontinuous beds, up to 3 m thick, close to the base of the valley fill near the Tagus confluence (Fig. 2). Its lower boundary with units 2 and 2a is gradual and conformable, whereas the contact with unit 1 is abrupt and conformable. Unit 4 (lithofacies Fp, C, Cs) is found in the central and upstream reaches of the lower valley fill (Fig. 2). The unit consists of discontinuous peat, silty peat and peaty silt beds and has a maximum thickness of 1.5 m. The habitat preferences of the species found within the peat bed of core 51 depict a freshwater marshy habitat with a silty substrate (Table 3). Organic material was degraded prior to burial and there are frequently large wood fragments present. The conformable lower boundary with units 2 and 3 is gradual to abrupt. Unit 5 (lithofacies Csx, Fpx) forms a distinct black bed in the upper part of the valley fill sequence which dips

sorting of unit 1a suggests that these sediments were deposited in, or near, narrow channels in a fine-grained floodplain. Unit 2 (lithofacies Fl, Sb-i) consists of discontinuous basal beds in the central and upstream reaches the finegrained valley fill with a maximum thickness of 2.5 m (Fig. 2). The unit comprises laminated sandy silts to bedded silty sands with frequent organic material and wood fragments (Fig. 3). Its conformable lower boundary with unit 1 is gradual. Sub-unit 2a (lithofacies Fle, Sb-ie) is found locally near the confluence with the Tagus River and has a maximum thickness of 0.7 m (Fig. 2). Its lithology is identical to unit 2; however, unit 2a contains also shells. The unit shares a gradual and conformable lower boundary with unit 1. Unit 3 (lithofacies Fo, Fom) forms the most extensive unit in the Muge valley fill (Fig. 2). One sub-unit (3a) is distinguished on the basis of shell presence. Units 3 and 3a form a thick fine-grained depositional wedge in the lower Muge valley fill, pinching out near the Lamarosa tributary (Fig. 1). The maximum thickness of the combined units is

Cumulative Loss on Frequency Ignition

Core 5

0

100%0

Loss on Ignition

Core 11

0

15%

0

15%

0

0

Cumulative Frequency

Core 20

Loss on Ignition Moisture

100% 0

20%

45%

A

M

M

Fx

2

20%

M

Fom

1

Fom

2

M

0

1

Fpx/Csx

M Fom Csx

2

M

C

Fx

M

100%

Fx

M

Fom Fp

Fpx/Fl C M

2

Fom

Cs

m

2

1

1

Fom Fpx Csx

Loss on Ignition

Core 64 0

M

m

1

M

0

0

C Fx

Loss on Ignition

Core 40

0

M

139

Fom

3

M

M

3

3

4

4

3

m

3

Fl

M 4

4

Fo 5

5

6

EOC/F

5

6

6

5

Sbi EOC/S C S Fs Ms Cs Gr

Fo

C S Fs Ms Cs Gr

5

m

C

4

Fom m

M

Fo

0 100 Grain Size (μm)

m below surface

Legend graphic logs

clay C S Fs Ms Cs Gr gravel silt coarse sand fine sand medium sand Cumulative Frequencies Clay (0-2 μm) Fine-Medium Silt (2-16 μm) Coarse-v.Coarse Silt (16-62 μm) Sand (>62 μm) Mean grain size (individual samples)

6

Sbi EOC/S C S Fs Ms Cs Gr

Fo 7

7

8

8

9

9

EOC/F C S Fs Ms Cs Gr

Other 0 100 Grain Size (μm)

Peaty Silty Peat

10

Wood fragments Foe Fle

11 11.5

Root/stem fragments

Sbie EOC/F C S Fs Ms Cs Gr

Peat Sedimentary Structures Lamination Traces of Lamination Traces of organic Lamination Gradational boundary Sharp planar boundary

Made ground

M

Nodules

Ploughed Root penetration Sandy inclusion

Mottling

C

Charcoal Leaf/plant fragmants

Shells and shell fragments Not recovered EOC/ End Of Core (on Sand/ S-F borehole failure)

Fig. 3. Sediment logs, grain size data and loss-on-ignition values for selected, representative cores. See Table 1 for lithofacies codes and Fig. 1 for location of cores.

ARTICLE IN PRESS T. van der Schriek et al. / Quaternary International 189 (2008) 135–151

140

Table 3 Preliminary results of plant macrofossil analysis of a peat sample at 2 m below the surface in core 51 (see Fig. 1 for location) Species

Number per 100 ml

Habitat

Scirpus cf. lacustris

4500

Hypericum tetrapterum (Syn. Hypericum acutum) Lycopus europaeus Juncus spp. Carex spp. Potamogeton spp. (degraded embryo)

4500

Wetlands, shallow pond and lakes and river margins, silt substrate, can survive water table variability and inundation Wetland, marshy

2 25 6 2

River banks, wetlands Wide niche, mainly damp areas Wide niches, mainly damp ground Aquatic

Plant material retained in a 125 mm sieve was scanned at up to 100  magnification for plant macrofossils under a reflected light microscope (analysed by Dr. J. Cotton).

towards the confluence with the Tagus River (Fig. 2). The unit forms a near-continuous layer in the downstream and central reaches, while discontinuous beds are present in the upstream reach. Thickness of this bed varies between 0.15 and 0.8 m. The unit comprises oxidised peaty clayey silt with lamination and cm-scale beds of grey inorganic silty clay, and high (up to 25%) loss-on-ignition values (Fig. 3). The sediments have a blocky structure and contain calcite concretions and abundant degraded organic material. Dark vertical stripes extend downprofile into units 2 and 3, probably representing root penetration. These characteristics are typical for an alluvial floodplain soil (USDA, 1975). Its conformable lower boundary is abrupt (units 2 and 3) to gradual (unit 4). Unit 6 (lithofacies Fom, Fx) caps the fine-grained lower Muge valley fill (Fig. 2). The unit has a maximum thickness of 3 m near the Tagus confluence and thins rapidly to 0.5–1.3 m upstream of coring site 20. The unit consists of reddish mottled clayey silt to sandy silt with occasional lamination and cm-scale sand beds, and contains frequent degraded organic material (Fig. 3). The lower boundary of unit 6 with units 2 and 5 is conformable and gradual. Where the unit overlies unit 3, it is not possible to define a lower boundary surface due to the homogenous nature of the sediments. 5. Chronology Two samples have been dated at the base of core 11: shell fragments (MUG-3) dating to 59107120 cal BC and plant fragments (MUG-4), which date to 61207100 cal BC and were sampled 7 cm above the shell sample (Table 2). There is a slight age reversal between the two dates, despite the age estimate on shell fragments being corrected for the marine reservoir effect. This age difference might be explained by the notoriously variable reservoir effect in

estuarine shells (e.g. Pilcher, 1991; Cearreta and Murray, 2000) and the fact that the dates have been obtained on different types of materials and calibrated against different curves. With these considerations taken into account, and on the basis of the overlap of the dates at the 2s range (Table 2), the difference between these age estimates is considered to be non-significant. The basal sample in core 20 (MUG-5) yielded an age of 6150790 cal BC, which is virtually identical to the basal date on plant fragments in core 11. Sample MUG-6 at 4.7 m depth in core 20 dates to 5555775 cal BC and aimed to inform the microfossil data (Section 6). These two age estimates are vertically consistent and correlate well with core 11 (Fig. 4); hence, they have been accepted. A bulk sample (MUG-2) of organic-rich clay at 3.6 m depth in core 20 yielded a date of 63257425 cal BC; this sample aimed to provide a range finding date (Passmore and Stevenson, 1999). The large error margins of this date suggest it to be an unreliable age estimate (Table 2). Furthermore, two lower dated samples (MUG-5 and MUG-6) in core 20 and the basal dates in core 11 are slightly younger (Fig. 4), suggesting that this radiometric date is too old. The problem may be caused by old (radioactive ‘‘dead’’) carbon that is present in estuarine and marine organic-rich mud (Section 6), and which renders radiocarbon age estimates older than the actual age of the dated sediment (e.g. Soter et al., 2001; Colman et al., 2002). Accordingly, this date is rejected. Sample MUG-7 at 2 m depth in core 20 yielded a date of 37957155 cal BC and aimed to inform the regional disappearance of estuarine environments, as indicated by pollen analyses (Section 7, PZ4). The date is consistent with the accepted chronology for the lower part of core 20 and with regional estuarine age–depth data (e.g. Zazo et al., 1994, 1996; Goy et al., 1996; Rodrı´ guez-Ramı´ rez et al., 1996; Morales, 1997; Borja et al., 1999; Dabrio et al., 1999, 2000; Psuty and Moreira, 2000; Cearreta et al., 2003; Freitas et al., 2003; Santos and Sa´nchez Gon˜i, 2003). Consequently, this date is accepted under the pretext that the radiocarbon age might be too old because of the presence of ‘‘dead’’ carbon in the analysed sample. The date of 48157125 cal BC on degraded wood fragments (MUG-8) from the basal buried soil level in core 20 (Table 2) is older than the lower date of 37957155 cal BC in core 20 (Fig. 4). Regional data indicate continued infilling of estuaries around 4800 cal BC, while published dates for soil formation in Atlantic Iberian estuaries date from 2000 cal BC onwards (Devoy et al., 1996; Goy et al., 1996; Granja and De Groot, 1996). The dating of reworked wood fragments, incorporated at the base of the alluvial soil, may have caused the older than expected age of the sample. Without further evidence supporting floodplain stabilisation and soil formation around 48157125 cal BC, this age estimate has been rejected. Finally, the date of 2307180 cal BC (MUG-1) in core 20, on a bulk sample from an individual peaty silt layer near the top of the buried soil, is consistent with the lower dates in core 20. Therefore, this date is accepted.

ARTICLE IN PRESS T. van der Schriek et al. / Quaternary International 189 (2008) 135–151

Fom

2

M

Scr obic ula ria Ca rdiu Plan a m Hy dro Edu le bia Ulv ae

M M Fom Csx

1 M Fom Fom

Csx

Csx

2

2 M

Csx M 3

M

3

Fom

Fom

3

M

M

Fx

m

M

M C

m

1

Fom

m

M

2

Fx

1

C

(c. 4m MdC)

Fx

Foraminifera Core 40 (no) (c. 5m MdC) Foraminifera 10 0 (No) Core 48 (c. 4.5 mMdC) 0 0 110 Fx (c. 4m MdC) Foraminifera 0 (no) M Fom 0 10 A 0 1 Fpx/ A Csx M Fom M Fx/Fpx 1 230+/-180 C Fpx/Fl Fpx cal BC C Csx Csx 1 4815+/-125 2 M Fom/ Fom M cal BC Fp Fom M 2 M 3795+/-155 2 3 cal BC Core 20

0

0 C

m

m

1

50

0 0

Fx

M

Foraminifera (No)

Core 13

m

0

(c. 4m MdC)

Core 12 (c. 4m MdC)

m

(c. 4m MdC)

Shell species

Shell species Core 11

Scr obic ula ria Ca rdiu Plan a m Hy dro Edu le bia Ulv ae

Core 10

Scr obic ula ria Ca rdiu Plan mE a Hy dro d bia ule U lv ae

Shell species

3

Fom 3

4 4

4

4

4

4

Fo

5

5

Fo

5

3

Fo

Fo

6325+/425 cal BC

5555+/75 cal BC

4

5

Fo 5

5

141

Sbi/Fl

β

Sbi

6

5

EOC/S

EOC/S 6 1240+/-160 cal BC 6

6 Fo

6

6

7

β

Fo

7

7

7

7

8

8

8

α

Foe

8 6150+/90 cal BC

8

Foe 9

9

9

Legend Microfossils

9 9

EOC/F

Fo

Calcareous foraminifera Agglutinating foraminifera

10 Foe

α

11 EOC/F

6120+/-100 cal BC 5910+/-120 cal BC

10 Foe Fle 11

Ostracod

Fl 10

EOC/S

α

10

Molluscs present

EOC/S ,

Biofacies assemblage

Sbie 11.5

EOC/F

Fig. 4. Graphic logs, foraminifera count diagrams and mollusc species of analysed cores. Fossil assemblages are explained in Section 6. Height of cores is given in metres above Mare´grafo de Cascais (Portuguese datum level). For location of cores see Fig. 1 and for graphic log legend see Fig. 3. Underlined radiocarbon dates are rejected (Table 2).

The basal contact of core 40, at 6 m below the surface, has been dated to 12407160 cal BC on plant fragments (MUG-9). This age estimate is anomalously young and inconsistent with the chronostratigraphic evidence of adjacent cores (Fig. 4). It is assumed that the sediments sampled for radiocarbon analysis were contaminated during coring and extraction in the field. Accordingly, this date is rejected.

accumulation rates during this period are likely to be greater than 0.26 mm/yr prior to soil formation, and lower than this rate during soil formation. However, the boundaries of the soil unit have not been reliably dated in core 20. From 230 cal BC onwards the average accumulation rate is 0.43 mm/yr. This rate is based on a date near the top of the soil, which probably causes a slight underestimation of the sediment accumulation rate of the overlying unit.

5.1. Sediment accumulation rates 6. Foraminifera, Ostracod and Mollusc analyses A sediment accumulation model has been constructed for core 20, based on linear extrapolation between the accepted radiocarbon dates (Fig. 5). The resulting curve displays the changing fine-grained sediment accumulation rates over time. The basal part of the sediment record shows a high accumulation rate of 7 mm/yr from 6150 to 5555 cal BC, contemporary with deposition in an estuarine environment (Section 6). Subsequently, the deposition rate slows down to 1.5 mm/yr from 5555 to 3795 cal BC, encompassing the change from estuarine to fluvial deposition at the site (Section 6). Between 3795 and 230 cal BC, the average sediment accumulation rate decreases to 0.26 mm/yr, corresponding with fluvial deposition and soil formation (Section 7). The actual

Ostracods and foraminifera have been analysed in cores 11, 20, 40 and 48, and molluscs in cores 10–13 (Fig. 4). The analyses aimed to characterise particular depositional environments rather than precise elevations relative to the tidal framework. Interpretation of foraminifera associations is based on established relationships between foraminifera assemblages and modern intertidal environments (e.g. Scott and Medioli, 1978; Cearreta, 1988, 1998; Boomer, 1998; Freitas et al., 1999; Haslett et al., 2001). Foraminifera sample weight varied from 1.5 to 4 g of sediment; species are displayed in raw counts as each sample had less than 250 specimens and is thus statistically not significant (Haslett et al., 2001). However, these data

ARTICLE IN PRESS T. van der Schriek et al. / Quaternary International 189 (2008) 135–151

142

0 Position and thickness of dated sample with 2σ error margin

1 2

6.1. Assemblage a

Depth in m below the surface

3 4 5 6 7 8 9 10 11 12 2000 1000

sediments, while sediment lamination was undisturbed and shells were frequently found in living (in-situ) position. In combination, these factors imply little or no erosion/reworking of the cored sediments.

Sample MUG 4

0

1000 2000 3000 4000 5000 6000 7000 8000 years AD/BC (calibrated)

Fig. 5. Sediment accumulation curve for core 20 in the lower Muge valley floor, based on the accepted radiocarbon dates (Table 2); the basal radiocarbon date on plant fragments in core 11 (MUG-4) has been included. For location of cores see Fig. 1.

are considered sufficiently accurate for the purpose of this study, which aims to characterise general depositional environments rather than exact sample position within the tidal framework. Therefore, larger error margins are acceptable. Furthermore, relatively low numbers of foraminifera are normal in saltmarsh environments (Boomer, 1998). Fig. 6 shows that the peaks in the raw foraminifera count curves of cores 11 and 20 are not an artefact of variances in sample weight used for the different analyses. Use of molluscs as qualitative palaeoenvironmental proxies was based on established relationships between modern molluscs and their habitat (Tucker Abbott, 1990; Peacock, 1993). Basal sediments, consisting of coarse sand and gravel (Unit 1; lithofacies Sb-ii, Sb-i), yielded no microfossils in the analysed cores. The well-sorted nature, bedding and coarse grain sizes of these sediments point towards deposition in a channel system. Samples from the finegrained valley fill that were prepared for foraminifera analyses contained organic inclusions (in particular insect remains and leaf fragments) and platy mica minerals that are indicative of very-low-energy depositional conditions. Abundant quantities of soft plant material, seeds, wood fragments and insect remains throughout the samples suggest a nearby terrestrial sediment source. There were no major erosive contacts present in the cored fine-grained

The basal parts of cores 10, 11, 12 and 13 all contain shells (lithofacies Foe, Fle and Sbie) which are often found in living position. Scrobicularia plana is by far the most abundant species in all of these cores, while Cardium edule/ exiguum shells are found in cores 10 and 11, and Hydrobia ulvae in core 10 (Fig. 4). The foraminifera assemblage in core 11, which overlaps with the shell beds at 11.0–10.2 m below the surface, is dominated by the calcareous species Haynesina germanica, Haynesina depresula and Ammonia beccarii. There is a single agglutinating specimen of Jadammina macrescens at 10.6 m below the surface (Fig. 4). These foraminifera are characteristic of a brackish, intertidal estuarine environment consisting of mudflats to lowmarsh. Furthermore, Scrobicularia plana and Cardium edule/exiguum shells are found 2–8 cm below the surface of muddy intertidal flats, whereas Hydrobia ulvae is found at the surface of the highest intertidal flats and saltmarshes. Accordingly, the combined fossil assemblage is deemed characteristic of (high) intertidal mudflats that are daily inundated by brackish water of variable salinity (Fig. 7). Two 14C dates at the base of core 11 establish the age of the shell-bearing facies at 6200 cal BC. 6.2. Assemblage b The second fossil assemblage contains no shells and is dominated by agglutinating foraminifera species. Agglutinating taxa Jadammina macrescens and Trochammina inflata are present in core 11 at 10.0–3.57 m below the surface, with peak values at 8.26 m. Between 7.26 and 5.74 m below the surface the assemblage contains also calcareous taxa Haynesina germanica, Haynesina depresula and Ammonia beccarii. This assemblage in core 11 dates from just after 6200 cal BC (Fig. 4). The foraminifera assemblage found in core 20 contains solely the agglutinating species Jadammina macrescens and Trochammina inflata between 8.7 and 3.7 m below the surface; peak values occur between 5.2 and 4.7 m (Fig. 4). The base of the core is dated at 6200 cal BC and the upper limb of the foraminifera peak at 5555 cal BC. The base of core 40 at 5.91 m below the surface contains low numbers of the agglutinating taxa Jadammina macrescens and Trochammina inflata. This basal foraminifera sample correlates stratigraphically with the foraminifera peak in core 20 (Fig. 4). There are single occurrences of Jadammina macrescens higher up in this core and in core 48. Mixed calcareous-agglutinating foraminifera assemblages containing Haynesina germanica, Haynesina depresula, Ammonia beccarii, Jadammina macrescens and

ARTICLE IN PRESS T. van der Schriek et al. / Quaternary International 189 (2008) 135–151

Depth below the surface (m)

0

Core 11: Total foraminifera 10 20 30 40

50

0

0

0

1

1

2

2

3

3

4

4

5

5

6

6

7

7

8

8

9

9

10

10

11

11 Core 20: Total foraminifera 0

Depth below the surface (m)

0 10 20 30 40 50 60 70 80 90 100110 0

0

1

1

2

2

3

3

4

4

5

5

6

6

7

7

8

8

9

9

Core 11: Species per gr 5 10 15

143

20

Core 20: Species per gr 100 200 300

400

Fig. 6. Foraminifera diagrams for cores 11 and 20 displaying raw count data and total number of species per gram. Peaks in the raw count- and species per gram curves match for each of the cores, indicating that raw count peaks are not an artefact of sample weight.

Trochammina inflata are characteristic for intertidal conditions in an upper mudflat to lowmarsh environment (Fig. 7). Three ostracods of the taxa Loxoconcha elliptica and Leptocythere porcellanea overlap with this mixed assemblage in core 11 and indicate the low-salinity end of brackish waters. Agglutinating foraminifera assemblages consisting solely of Jadammina macrescens and Trochammina inflata are characteristic of the upper intertidal range in brackish high- to uppermarsh environments (Fig. 7). Relatively low numbers of foraminifera are normal in saltmarsh environments, which show increasingly lower abundances towards the upper margin of the tidal range (Boomer, 1998).

6.3. Local estuarine development Foraminifera and shell analyses show that the lower Muge valley near the Tagus confluence experienced an abrupt beginning of fine-grained sedimentation in an estuarine intertidal setting, indicating a sudden drowning of the valley floor around 6200 cal BC. The presence of basal shell-rich beds throughout cores 10, 11 and 12 suggests that the initial saltwater intrusion in the Muge valley created extensive mudflats near the confluence with the Tagus River. In coring sites 10 and 11, there follows within 1.3 m from the base of the valley fill a transition from mudflat to saltmarsh environments, indicating a rise

ARTICLE IN PRESS T. van der Schriek et al. / Quaternary International 189 (2008) 135–151

amm ina

Local Zonation

rToch

Ja d

Tide Levels

m ac resc ens amm ina Am infla mon ta ia b ec c arii Hay nes ina germ an d e p i ca / Scr r e su obic la ular ia p lana Car dium edu le/e Hyd xigu robi um a ul v ea

144

Environmental Characteristics Freshwater

Local Environment Alluvial Floodplain

HAT

Barren

MHWST

J. macrescens and T. inflata No molluscs

Rarely tidally flooded, low salinity. Mud

Upper-High Saltmarsh

J. macrescens and T. inflata, occ. A. beccarrii No molluscs

Periodically tidally flooded, brackish. Mud

Middle-High Saltmarsh

MHW

H. germanica and A. beccarii MHWNT

Rare molluscs

High intertidal, regularly flooded, brackish. Mud-sandy mud

Low Saltmarsh

Lowmarsh-Mudflat LHWNT

H. germanica

MTL

Molluscs dominated by S. plana

Intertidal, diurnal flooded, brackish. Mud-sandy mud

Mudflat

Fig. 7. Summary diagram of depositional environments of foraminifera and shell groups in the lower Muge valley fill (Section 6). Local zonation, environmental characteristics and depositional environment are based on foraminifera (Scott and Medioli, 1978; Cearreta, 1988, 1998; Boomer, 1998; Freitas et al., 1999; Haslett et al., 2001) and shell habitat (Barret and Younge, 1958; Tucker Abbott, 1990; Peacock, 1993). Dashed lines indicate the tidal range in which species are rare (HAT: Highest Astronomical Tide; MHWST: Mean High Water Spring Tide; MHW: Mean High Water; MHWNT: Mean High Water Neap Tide; LHWNT: Lowest High Water Neap Tide; MTL: Mean Tidal Level).

of these sites within the tidal framework. However, stratigraphic cross-correlation reveals that upstream coring sites 12 and 13 continued to support mudflats higher up in the valley fill record (Fig. 4). This may suggest that mudflat environments became confined to the margins of the tidal Muge River channel. In the vicinity of coring sites 13, 20 and 40 (Fig. 1), the initiation of fine-grained sedimentation started in marginal tidal environments. At the locality of core 13, tidal and saltwater influence grew until a local maximum was reached and estuarine shell beds formed in a mudflat environment, 2 m above the basal contact. Upstream of core 13, shells disappear from the lower Muge valley fill (Fig. 2). Fine-grained sedimentation near coring site 20 started 6200 cal BC in a marginal saltmarsh environment. Tidal influence increased at this locality and the peak in foraminifera numbers between 5.2 and 4.7 m probably indicates the period of maximum saltwater and tidal influence. Linear extrapolation between dated samples in core 20 (Fig. 5) suggests that the lower limb of this peak dates to 5800 cal BC; the upper limb dates to 5555 cal BC. This period was followed by a decrease in saltwater influence and renewed marginal tidal conditions. Tidal influence of this locality ends 3.2 m below the surface based on disappearance of foraminifera; linear

extrapolation between dated samples suggests an age of 4700 cal BC for this depth (Fig. 5). Maximum inland saltwater penetration is recorded at the base of core 40, 4.5 km upstream from the Tagus confluence, where fine-grained sedimentation started in a marginal tidal environment. The base of core 40 has a (rejected) radiocarbon date of 12407160 cal BC (MUG-9; Table 2). Pollen analyses show that regional saltwater indicators are lost after 3795 cal BC (Section 7), which suggests strongly that the basal foraminifera were not deposited at 12407160 cal BC. Stratigraphic cross-correlation of the base of core 40 with core 20 suggests that tidal influence at the base of core 40 dates to 5555 cal BC (Fig. 4). The presence of individual agglutinating foraminifera in the lower parts of cores 40 and 48 may point to occasional flooding by saltwater. 7. Vegetation history Core 20, located in the Muge valley floor midway between the archaeological sites of Moita do Sebastia˜o and Cabec- o da Amoreira (Fig. 1), has been analysed in detail for pollen. The summary pollen diagram is confined to the most significant pollen types and includes the raw foraminifera count data for core 20 (Fig. 8). The fine-

ARTICLE IN PRESS T. van der Schriek et al. / Quaternary International 189 (2008) 135–151

7.1. Pollen zone PZ1 The basal pollen zone (PZ 1) spans sediments (lithofacies Fo) between 9.00 and 6.85 m depth and was deposited after 6200 cal BC. This zone is dominated by significant quantities of pine and deciduous oak pollen. The base of the zone has low values for pine and juniper and relatively high values for weeds (e.g. Brassicaceae, Asteraceae and Poaceae), which probably indicate an open or semi-open landscape. Subsequently, oak and particularly pine pollen values increase, although the presence of Erica arborea, Calluna, Genista (Fabaceae), Lamiaceae and Cistus ladanifer (a fire indicator) point towards an open woodland environment. Pine probably occupied the free-draining sandy soils adjacent to the valley floor, while semideciduous oak Q. faginea would have occupied the fringes of the freshwater valley floor and gullies with more moisture-rich soils. Occasional individuals of Jadammina and Trochammina in the valley fill sediments of PZ 1 indicate that the site was at the margins of tidal influence. The presence of Chenopodiaceae pollen, a saltmarsh indicator, in this zone supports the interpretation of a brackish water influenced local valley floor. Isoetes spores (probably I. histrix) follow the pattern of Chenopodiaceae values, and it has been suggested that this species might also indicate brackish marsh environments in the Mediterranean (Stevenson and Battarbee, 1991; Stevenson and Harrison, 1992). This species probably occurred in reduced salinity conditions around the Highest Astronomical Tide (HAT) level where it was not regularly flooded by brackish water. Freshwater indicators such as Alnus, Salix, Cyperaceae, Ranunculus and Equisetum suggest that upstream parts of the valley floor, in close proximity to the coring site, lay beyond the limit of the saltwater intrusion and were characterised by alder woodland and freshwater marsh. 7.2. Pollen zone PZ2 Subsequently, in zone PZ 2 between 6.75 and 4.00 m depth (lithofacies Fo), increasing quantities of Chenopodiaceae pollen, together with a peak in numbers of agglutinating foraminifera, indicate that the Muge valley floor was tidally influenced with brackish water and extensive areas of high saltmarsh at the time of local Mesolithic occupation (6200–4800 cal BC). The age of the lower limb of the foraminifera peak at 5.2 m is inferred at 5800 cal BC (Section 6.3), while the upper limb (at a depth

Fig. 8. Summary pollen diagram of core 20 in the lower Muge valley floor (analysed by F. Franco Mugica) including raw foraminifera counts. Dates are given in years cal BC and underlined dates are rejected (Table 2). For legend of graphic log see Fig. 3.

grained valley fill is most likely predominantly derived from the upstream Muge catchment; however, the lower part of the core was tidally influenced which implies a potential tidal sediment input from the lower Tagus catchment. Accordingly, the pollen source area was presumably a combination of local and regional inputs. To date no studies for the western Mediterranean have quantified the relative size of these inputs.

145

ARTICLE IN PRESS 146

T. van der Schriek et al. / Quaternary International 189 (2008) 135–151

of 4.7 m) is dated to 5555 cal BC. Isoetes spores, Liguliflorae and Cyperaceae pollen also increase through this zone and suggest that the fringes of the valley floor consisted of marshland at, or above, the HAT. Alnus pollen indicate that there is still freshwater vegetation in upstream parts of the basin during this period, although the decreased percentages might reflect the poor saline water tolerance of this species. Regional vegetation in the catchment sees the maintenance of oak forest, but the pine woodland suffers progressive losses that are interspersed with regeneration episodes. Declines in pine pollen are associated with increases in herbs such as Liguliflorae and especially Brassicaceae pollen that indicate open spaces in the landscape. It is interesting to note that three phases of cereal-type pollen appear in both this and the previous zone. 7.3. Pollen zone PZ3 At the start of zone PZ 3 (4.00–2.00 m below the surface; lithofacies Fo and Fom), the regional vegetation sees a significant diminution in the extent of pine woodlands. Pine pollen remains at its lowest observed level over zone PZ3 and, together with the sustained high values for herbs (e.g. Stachys t., Astragalus and Brassicaceae), suggests that the regional landscape experienced sustained deforestation. The deposition of PZ3 was associated with declining sediment accumulation rates. Hence, low Pine pollen values in this zone most probably reflect low pollen production rates as there was no increase in the accumulation speed to account for the low numbers. At present, we have no radiocarbon sample dating the Pine decline. However, linear extrapolation between dated samples (Fig. 5) suggests that this event at 4.00 m depth dates to 5000 cal BC. Van Leeuwaarden and Janssen (1985) inferred a similar regional decline in pine woodland, predating 5000 BP (3800 cal BC), near Alpiarc- a, located 19 km north of the Muge confluence. A regional decline in Pine woodland is also registered in the Alentejo region of Southern Portugal around 6560 BP (5500 cal BC) at Lagoa Travessa and 5300 BP (4200 cal BC) at Ribeira de Moinhos, respectively 90 and 210 km south of Muge village (Mateus and Queiroz, 2000). At the Santo Andre´ coastal area, some 150 km south of Muge, Pinus forests suffer a major decline 5300 BP (4200 cal BC) (Santos and Sa´nchez Gon˜i, 2003). At these sites, the replacement of pine forests by evergreen Quercus, while deciduous Quercus pollen percentages decrease, may suggest a regional drying trend around 5500–4000 cal BC. The Muge valley adjacent to the Amoreira shell midden appears to have lost most of its saltwater influence during the period covered by this zone. Jadammina and Trochammina values decline to zero above a depth of 3.2 m, inferred to correspond with an age of 4700 cal BC (Section 6.3), whereas marsh indicators such as Liguliflorae and Isoetes increase. This vegetation is not regularly flooded by brackish water and is located at, or above,

HAT. However, Chenopodiaceae values increase towards their maximum value of 60%, indicating significant quantities of saltmarsh in close proximity to the core site. Alnus values remain similar to the previous zone, suggesting an open, wet floodplain environment in the local valley floor with wooded patches in the upstream area. Traces of (open) freshwater taxa increase towards the upper part of this zone, notably Typha, Myriophyllum alterniflorum and Nymphaea that are indicative of standing or low-energy waters. At the start of the freshwater transition of the tributary valley floor, the local Mesolithic record ends (4800 cal BC) and the area is abandoned. 7.4. Pollen zone PZ4 Over pollen zone PZ 4, between 2.00 and 1.00 m (lithofacies Fom, Csx and Fpx), the local valley floor environment appears to change significantly as all indications of saline conditions (e.g. Chenopodiaceae and Isoetes) are abruptly lost near the base of the zone (3800 cal BC). These species are replaced by freshwater indicators, notably Myriophyllum alterniflorum, Typha and Nymphaea that indicate the presence of still, shallow open water. The open-water communities in the lower part of this zone are progressively invaded by reeds (Filicales) and, especially, by alder woodland which reaches maximum values prior to 230 cal BC at the base of the buried soil (lithofacies Csx). Subsequently, Alnus pollen values decrease and cereal-type pollen are present in the upper part of this zone. Regional tree cover appears to experience a small resurgence of pine, with a corresponding loss of oak woodlands. Their replacement by prominent shrub communities dominated by Erica arborea, Calluna, Cistus, Genista and Helianthemum and ruderals like Fabaceae, Rumex or Plantago indicate major anthropogenic disturbance of catchment vegetation. There is no archaeological evidence of occupation of the area after Mesolithic site abandonment up to the Roman period; however, isolated archaeological finds reveal a potential late Neolithic, Bronze- and Iron Age presence (Cruz and Oosterbeek, 1993; Lucas and Ferrari, 1993; Joacquinto, pers. comm.) which appears to be supported by the pollen data showing major human impact upon the catchment vegetation after 3800 cal BC. 7.5. Pollen zone PZ5 Between 1.25 and 0.50 m (lithofacies Fom), overlapping with the end of pollen zone PZ 4 and zone PZ 5, the conversion of the valley for cereal cultivation is seen following removal of the local alder woodland (230 cal BC). This transition coincides with the lithological transition from dark, peaty silt (lithofacies Fpx) to grey clayey silt (lithofacies Fom), signifying soil burial. Occasional high values for Myriophyllum alterniflorum and Typha throughout zone PZ 5 suggests that parts of the valley floor continued to support open-water conditions such as shallow lakes and ponds. Preliminary diatom

ARTICLE IN PRESS T. van der Schriek et al. / Quaternary International 189 (2008) 135–151

analysis of one sample from a white layer in the buried soil level (core 40; lithofacies Fpx/Csx) supports this interpretation. The sample consisted almost entirely of diatoms and yielded species indicative of an environment of shallow standing water, while the lack of minerogenic material argues against the sample being a flood deposit (Clarke, pers. comm.). The dominant genera are Eunotia and Tabellaria fenestrate; Cyclotella meneghiniana, Cymbella spp., Fragilaria spp., Pinnularia spp., Nitzschia frustulum, Achnanthes, Navicula, Achnanthidium and Aulacoseira spp. make up the rest of the assemblage. Shrub communities indicated by Erica arborea and Calluna decrease and regional tree cover consists for the first time of substantial amounts of Olea. It is possible that the shrub communities, which appeared after deforestation of the catchment area, were removed in order to obtain open spaces for olive cultivation. Although the area remains virtually deforested, the slight increase in total pollen concentration suggests a more extensive and mainly herbaceous vegetation cover. The first recorded evidence for valley floor cultivation and production of olives after 230 cal BC coincides with archaeological evidence showing the presence of Roman buildings in the Muge floodplain and the lower valley (Batata and Gaspar, 1993). The early Medieval period is poorly represented in the archaeological record. Historic evidence suggests that the Muge region became an important cereal producing area prior to the 13th century (Batata and Gaspar, 1993; Vilar, 1993). 8. Holocene environmental change in the lower Tagus valley 8.1. Vegetation development The critical factor influencing vegetation changes in western Iberia is believed to be precipitation, while temperature changes are of secondary importance (Allen et al., 1996; Carrion, 2002). During the late glacial period, the Estremadura hills bordering the Atlantic Ocean in central Portugal supported open woodland dominated by Pinus sylvestris, Buxus sempervirens and Leguminosae and served as refugia for thermophilous taxa such as Quercus (deciduous), Olea europaea and Arbutus unedo (Figueiral and Terral, 2002). There was rapid reforestation after the late glacial maximum and a well-developed Pinus forest appeared on the interfluves in the south-western Portuguese coastal area, while oakwoods occupied the valleys (Mateus and Queiroz, 2000; Santos and Sa´nchez Gon˜i, 2003). These woodlands existed throughout the Younger Dryas, attesting to a relatively mild climate. From 9500 to 4200 cal BC, the coastal area was dominated by a Mediterranean forest with Pinus (Mateus, 1985; Mateus and Queiroz, 2000; Santos and Sa´nchez Gon˜i, 2003). Xerothermic forest dominated by Quercus developed after 9500 cal BC in the central Portuguese mountains and changed after 7600 cal BC to mesothermic forest, indicating slightly cooler conditions (van der Knaap and van Leeuwen, 1994, 1995, 1997). Mediterranean taxa domi-

147

nated the Atlantic hills of the Estremadura from the early Holocene period, and average temperatures and precipitation were comparable to the present day throughout the Holocene (Figueiral and Terral, 2002). The vegetation record of core 20 in the lower Muge valley, beginning at 6200 cal BC, indicates a relatively dry climate in the central Portuguese lower Tagus basin, located in the rain shadow of the Atlantic Estremadura hills. Initially, there was an open woodland environment dominated by Pine on the interfluves, while semi-deciduous oak Q. faginea probably occupied the valleys. The open Pine woodland environment declined progressively and an open landscape was established around 5000 cal BC on the interfluves in the lower Tagus area. Several other pollen sequences in southern Iberia show a decline of Pine woodlands, suggesting a regional drying trend between 5500 and 4000 cal BC (e.g. Mateus, 1985; Van Leeuwaarden and Janssen, 1985; Carrion and Dupre´, 1996; Carrion and Van Geel, 1999; Yll et al., 1999; Mateus and Queiroz, 2000; Carrion et al., 2001; Santos and Sa´nchez Gon˜i, 2003). This drying trend may relate to the fall in lake levels and increased aeolian activity recorded after 3000 cal BC in southern Iberia (Harrison and Digerfeldt, 1993; Taylor et al., 1998; Borja et al., 1999; Carrion and Van Geel, 1999; Zazo et al., 1999; Carrion et al., 2001; Reed et al., 2001; Carrion, 2002). Central Portuguese vegetation records indicate small scale human-induced deforestation from the Chalcolithic (4600 cal BC) onwards, which intensified around the Bronze Age (1600 cal BC) and led to large scale deforestation in Roman times, followed by forest regeneration during the early Medieval period (van Leeuwaarden and Janssen, 1985; Van der Knaap and van Leeuwen, 1994, 1995). The most extensive human impact on the Portuguese landscape took place from the 12th century onwards and led to the virtual disappearance of forests in the 15th century (Mateus, 1985; van Leeuwaarden and Janssen, 1985; Van der Knaap and van Leeuwen, 1995; Mateus and Queiroz, 2000; Santos and Sa´nchez Gon˜i, 2003). Human disturbance of the Muge catchment vegetation started after 3800 cal BC and was initially confined to the interfluves. More intensive human impact on the vegetation is registered from 230 cal BC, when the valley floor was partially conversed for cereal cultivation and Olea cultivation was introduced to the interfluves. Historical records suggests that the Muge area was an important cereal producing area prior to the 13th century, although the upstream parts of the lower valley floor were only cultivated after 1850 (Batata and Gaspar, 1993; Vilar, 1993). 8.2. Estuarine evolution A review study of Portuguese relative sea-level (RSL) data by Dias et al. (2000) presents a single sea-level band, showing no data points with their error margins. This curve indicates that RSL rose rapidly from the last glacial

ARTICLE IN PRESS 148

T. van der Schriek et al. / Quaternary International 189 (2008) 135–151

maximum until 7000 cal BC (8000 BP), when RSL was 20 m below present levels. There followed a slowdown in the rate of RSL rise until 2600 cal BC (4000 BP) when sea level stabilised. Psuty and Moreira (2000) have suggested a low rate of sea-level rise in the Sado estuary after 2600 cal BC up to the present. However, southern Spanish RSL investigations, reviewed by Zazo et al. (1996), indicate stable sea levels at the Atlantic Iberian coast around 6300 cal BP (7500 BP). The rapid rate of early Holocene RSL rise outstripped the rate of sediment accumulation in estuaries across Atlantic Iberia (e.g. Goy et al., 1996; Rodrı´ guez-Ramı´ rez et al., 1996; Morales, 1997; Borrego et al., 1999; Dabrio et al., 2000; Lobo et al., 2001; Boski et al., 2002). The earliest estuarine deposits in valley fills along the south Iberian Atlantic coastline date around 9800 cal BC and testify to a transformation from brackish to open marine systems as RSL rose (e.g. Dabrio et al., 2000; Boski et al., 2002; Freitas et al., 2003). Extensive infilling of the southern Iberian estuaries started 5900 cal BC, when the rate of sediment supply began to balance the decreasing rates of sea-level rise (Dabrio et al., 1999, 2000; Boski et al., 2002). The beginning of rapid fine-grained sedimentation in the lower Muge valley floor around 6200 cal BC coincided with the establishment of estuarine environments in the lower 3.5 km of the tributary floodplain. This date for the onset of fine-grained estuarine sedimentation is consistent with published age/depth relationships for estuarine records in Southern Portugal (Mateus, 1985; Psuty and Moreira, 2000; Cearreta et al., 2003; Freitas et al., 2003). The date for the beginning of Mesolithic occupation of the area (6200 cal BC) overlaps with the age for the establishment of estuarine environments in the tributary. Mesolithic communities, largely sustained by estuarine resources, were probably attracted by the rich and variable range of saltwater, freshwater and terrestrial environments in immediate proximity to the sites. Tidal influence of the lower Muge floodplain increased up to 5800–5500 cal BC, when estuarine environments reached their greatest extent. This date is consistent with other Atlantic Iberian estuarine records registering a maximum transgressive surface around 5700–4100 cal BC (Mateus, 1985; Zazo et al., 1994, 1996; Goy et al., 1996; Rodrı´ guez-Ramı´ rez et al., 1996; Morales, 1997; Borja et al., 1999; Dabrio et al., 1999, 2000; Psuty and Moreira, 2000; Cearreta et al., 2003; Freitas et al., 2003; Santos and Sa´nchez Gon˜i, 2003). There followed a contraction of estuarine environments in the lower Muge valley from 5500 cal BC, until tidal conditions abruptly disappeared around 3800 cal BC. The fluvial conversion of the tributary valley floor is associated with deteriorating drainage conditions, as standing shallow water bodies appeared for the first time in the floodplain. The sediment accumulation curve (Fig. 5) indicates that aggradation rates decreased over this period, which implies that the saltto freshwater transition was not linked to an increase in the rate of sediment supply. The fine-grained fluvial sediments

corresponding to the top of unit 3 (Fig. 4) are capped by an alluvial soil, implying low rates of clastic aggradation and therefore a stable base- and sea level. There is limited evidence of soil formation in Atlantic Iberian estuaries and coastal settings dating to 2000–1400 cal BC (Devoy et al., 1996; Goy et al., 1996; Granja and De Groot, 1996). The rapid mid-Holocene infill of estuarine accommodation space is commonly linked to the slowdown in eustatic sealevel rise, which shifted the balance between the rates of sea-level rise and sediment supply dramatically (Long, 2001). During the late Holocene, human disturbance of catchments becomes an important factor controlling the sediment input into estuaries (e.g. Freitas et al., 1999, 2003; Morhange et al., 1999; Dabrio et al., 2000; Dias et al., 2000; Boski et al., 2002), although the extent of human impact on fluvial sediment supply to estuaries remains a source of contention (e.g. Bintliff, 2002). Renewed finegrained alluvial sedimentation starts after 230 cal BC near the mouth of the lower Muge valley floor. The surface of the buried soil-level downstream of core 20 is located up to 2 m below present Tagus River level (Fig. 2). Renewed late Holocene aggradation in the Muge confluence zone was most likely related to internal mechanisms such as progressive fluvial infill of the lowland valley under influence of stable sea levels, or to external factors such as human- and/or climate controlled increases in the rate of fluvial sediment supply. Upstream of core 20 in the lower Muge valley, the buried soil surface is located above the base level of the Tagus River and local alluviation probably reflects an enhanced sediment supply to the Muge River, due to intensive post-medieval human disturbance of the catchment vegetation (Batata and Gaspar, 1993; Vilar, 1993). 9. Conclusions This study presents the first detailed record of Holocene environmental change in the lower Tagus area (Portugal), based on the fine-grained lower Muge valley fill. The pollen record of core 20 in the lower Muge valley indicates a relatively dry climate for the lower Tagus area since 6200 cal BC. An open woodland environment dominated by Pine, present from 6150 cal BC, is replaced around 5000 cal BC by an open landscape with Oak, reflecting a regional drying trend. From 3800 cal BC there is major disturbance of the catchment vegetation attributed to human activity. Human impact on the natural vegetation increased around 230 cal BC in the Muge area and involved clearance of the floodplain vegetation, and intensive cereal and Olive cultivation. Deposition of the fine-grained Holocene lower Muge valley fill was principally controlled by sea-level change; human impact on the regional catchment vegetation only gained importance as a factor controlling infill during the late Holocene. Fine-grained sediments started to backfill the lower Muge valley floor from 6200 cal BC, forced by

ARTICLE IN PRESS T. van der Schriek et al. / Quaternary International 189 (2008) 135–151

RSL rise. From 6200 to 3800 cal BC, the lower Muge valley was occupied by an inner estuary at the upper limit of tidal influence. The shifting boundary between tidal and freshwater floodplain environments in the lower Muge valley floor was a function of the inherited valley floor topography and the balance between the rates of RSL rise (creating accommodation space) and sediment input. Estuarine environments included mudflats and low-, high, and upper saltmarsh, all of which are found in the upper tidal framework (Fig. 7). The similar position in the tidal framework occupied by estuarine environments in the lower Muge valley floor throughout the period of saltwater influence suggests that the rate of sediment accumulation broadly kept pace with the rate of sea-level rise. Initially, estuarine environments merged into an alluvial floodplain 3.5 km upstream from the Tagus confluence. During the period of maximum tidal influence of the lower Muge floodplain (5800–5500 cal BC) the boundary with the freshwater floodplain was located 4.5 km upstream from the Tagus confluence. Subsequent contraction and final disappearance (3800 cal BC) of estuarine environments has been linked to the slowdown in mid-Holocene eustatic sea-level rise, which promoted infill of the estuarine accommodation space. Stabilisation of the alluvial lower Muge floodplain and extensive soil formation prior to 230 cal BC has been linked to stable sea- and base levels. Renewed localised deposition from 230 cal BC was most likely related to internal mechanisms, such as progressive fluvial infill under influence of stable sea levels, or to external factors such as human- and/or climate controlled fluvial sediment supply increases. Mesolithic shell midden cultures in the lower Muge valley occupied a tidally influenced, aggrading valley floor that lies buried beneath the presently drained, deforested and cultivated freshwater floodplain. The establishment of estuarine conditions in the lower valley floor around 6200 cal BC appears to overlap with the beginning of Mesolithic settlement on the adjacent terrace levels. Mesolithic site abandonment around 4800 cal BC coincided with gradual environmental changes such as the establishment of a dryer climate and an open landscape in the lower Tagus area, while estuarine environments contracted slowly from 5500 cal BC. However, estuarine conditions near the mouth of the lower Muge valley and in the adjacent lower Tagus floodplain existed up to 3800 cal BC, at least 1000 years after site abandonment. Accepted archaeological hypotheses linking Mesolithic settlement abandonment to disappearing estuarine environments and resources need to be revised in the light of these new data. Acknowledgements The authors wish to thank the Camara Municipal de Salvaterra de Magos for much appreciated logistical support and Anabela Joaquinto for her invaluable assistance. We also thank Dona Teresa (Casa Cadaval) for fieldwork permission and support. This research was

149

undertaken while in receipt of a studentship of the Department of Geography, University of Newcastle upon Tyne. The seven AMS radiocarbon dates cited in this paper were granted by the NERC radiocarbon laboratory (radiocarbon dating allocation No. 923.0501): their support is gratefully acknowledged. Finally, we would like to thank referees Dr. Alessandro Arnorosi and Dr. Maria Fernanda Sa´nchez Gon˜i for their helpful and constructive comments. References Allen, J.R., Huntley, B., Watts, W.A., 1996. The vegetation and climate of northwest Iberia over the last 14,000 years. Journal of Quaternary Science 11, 125–148. Barbosa, B.P., 1995. Alostratigrafia e litostratigrafia das unidades continentais da Bacia Terciaria do Baixo Tejo-Relac- o´es com o eustatismo e a tectonica. Unpublished Ph.D. Thesis, Universidade de Lisboa, Lisboa. Barret, J., Younge, C.M., 1958. Collins Pocket Guide to the Sea Shore. Collins, London. Batata, C., Gaspar, F., 1993. Novos dados sobre a estac- a˜o arqueolo´gica de Porto do Sabugueiro (Muge, Salvaterra de Magos). O Foral, Vol. 1, pp. 24–33. Battarbee, R.W., 1986. Diatom analysis. In: Berglund, B.E. (Ed.), Handbook of Holocene Palaeoecology and Palaeohydrology. Wiley, Chichester, pp. 527–570. Bicho, N., 1994. The end of the Paleolithic and the Mesolithic in Portugal. Current Anthropology 35, 664–674. Bintliff, J., 2002. Time, process and catastrophism in the study of Mediterranean alluvial history: a review. World Archaeology 33 (3), 417–435. Boomer, I., 1998. The relationship between meiofauna (Ostracoda, Foraminifera) and tidal levels in modern intertidal environments of North Norfolk: a tool for palaeoenvironmental reconstruction. Bulletin of the Geological Society of Norfolk 46, 17–29. Borja, F., Zazo, C., Dabrio, C.J., Diaz del Olmo, F., Goy, J.L., Lario, J., 1999. Holocene aeolian phases and human settlements along the Atlantic coast of southern Spain. The Holocene 9 (3), 333–339. Borrego, J., Ruiz, F., Gonzalez-Regalado, M.L., Pendon, J.G., Morales, J.A., 1999. The Holocene transgression into the estuarine central basin of the Odiel River mouth (Cadiz gulf, SW Spain): lithology and faunal assemblages. Quaternary Science Reviews 18, 769–788. Boski, T., Moura, D., Veiga-Pires, C., Camacho, S., Duarte, D., Scott, D.B., Fernandes, S.G., 2002. Postglacial sea-level rise and sedimentary response in the Guadiana Estuary, Portugal/Spain border. Sedimentary Geology 150, 103–122. Brasier, M.D., 1980. Microfossils. George Allen and Unwin, London, 193pp. Bronk Ramsey, C., 1995. Radiocarbon Calibration and Analysis of Stratigraphy: The OxCal Program. Radiocarbon 37 (2), 425–430. Cabral, J., 1995. Neotecto´nica em Portugal Continental. Memo´rias, 31. Instituto Geolo´gico e Mineiro, Lisboa, 260pp. Carrion, J.S., 2002. Patterns and processes of Late Quaternary environmental change in a montane region of southwestern Europe. Quaternary Science Reviews 21, 2047–2066. Carrion, J.S., Dupre´, M., 1996. Late Quaternary vegetational history at Navarre´s, eastern Spain. A two core approach. New Phytologist 134, 177–191. Carrion, J.S., Van Geel, B., 1999. Fine-resolution Upper Weichselian and Holocene palynological record from Navarre´s (Valencia, Spain) and a discussion about factors of Mediterranean forest succession. Review of Palaeobotany and Palynology 106, 209–236. Carrion, J.S., Andrade, A., Bennett, K.D., Navarro, C., Munuera, M., 2001. Crossing forest thresholds: inertia and collapse in a Holocene sequence from south-central Spain. The Holocene 11 (6), 635–653.

ARTICLE IN PRESS 150

T. van der Schriek et al. / Quaternary International 189 (2008) 135–151

Cearreta, A., 1988. Distribution and ecology of benthic foraminifera in the Santon˜a estuary, Spain. Revista Espan˜ola de Paleontologı´ a 3, 23–38. Cearreta, A., 1998. Holocene sea-level change in the Bilbao estuary (north Spain): foraminiferal evidence. Micropaleontology 44 (3), 265–276. Cearreta, A., Murray, J.W., 2000. AMS 14C dating of Holocene estuarine deposits: consequences of high-energy and reworked foraminifera. The Holocene 10 (1), 155–159. Cearreta, A., Cacha˜o, M., Cabral, M.C., Bao, R., Jesus Ramalho, M.d., 2003. Lateglacial and Holocene environmental changes in Portuguese coastal lagoons 2: microfossil multiproxy reconstruction of the Santo Andre´ coastal area. The Holocene 13 (3), 447–458. Colman, S.M., Baucom, P.C., Bratton, J.F., Cronin, T.M., McGeehin, J.P., Willard, D., Zimmerman, A.R., Vogt, P.R., 2002. Radiocarbon dating, chronologic framework, and changes in accumulation rates of holocene estuarine sediments from Chesapeake Bay. Quaternary Research 57, 58–70. Cruz, A.R., Oosterbeek, L.M., 1993. A Arqueologia de Salvaterra. O Foral, 5, Vol. 5. Cunha, E., Cardoso, F., 2003. New data on muge shell middens: A contribution to more accurate numbers and dates. Muge Estudos Arqueolo´gicos 1, 171–184. Dabrio, C.J., Zazo, C., Lario, J., Goy, J.L., Sierro, F.J., Borja, F., Gonzalez, J.A., Flores, J.A., 1999. Sequence stratigraphy of Holocene incised-valley fills and coastal evolution in the Gulf of Cadiz (southern Spain). Geologie en Mijnbouw 77 (3–4), 263–281. Dabrio, C.J., Zazo, C., Goy, J.L., Sierro, F.J., Borja, F., Lario, J., Gonza´lez, J.A., Flores, J.A., 2000. Depositional history of estuarine infill during the last postglacial transgression (Gulf of Cadiz, Southern Spain). Marine Geology 162, 381–404. Detry, C., 2001. Arqueozoologia em concheiros de Muge. Unpublished M.Sc. Thesis, Autonomous University of Lisbon, Lisbon, 97pp. Devoy, R.J.N., Delaney, C., Carter, R.W.G., Jennings, S.C., 1996. Coastal stratigraphies as indicators of environmental changes upon European Atlantic coasts in the Late Holocene. Journal of Coastal Research 12 (3), 564–588. Dias, J.M.A., Boski, T., Rodrigues, A., Magalha˜es, F., 2000. Coast line evolution in Portugal since the Last Glacial Maximum until present—a synthesis. Marine Geology 170, 177–186. Figueiral, I., Terral, J.-F., 2002. Late Quaternary refugia of Mediterranean taxa in the Portuguese Estremadura. Quaternary Science Reviews 21, 549–558. Freitas, M.C., Andrade, C., Moreno, J.C., Munha, J.M., Cachao, M., 1999. The sedimentary record of recent (last 500 years) environmental changes in the Seixal Bay marsh, Tagus estuary, Portugal. Geologie en Mijnbouw 77 (3–4), 283–293. Freitas, M.d.C., Andrade, C., Rocha, F., Tassinari, C., Munha´, J.M., Cruces, A., Vidinha, J., Marques da Silva, C., 2003. Lateglacial and Holocene environmental changes in Portuguese coastal lagoons 1: the sedimentological and geochemical records of the Santo Andre´ coastal area. The Holocene 13 (3), 433–446. Goy, J.L., Zazo, C., Dabrio, C.J., Lario, J., Borja, F., Sierro, F.J., Flores, J.A., 1996. Global and regional factors controlling changes of coastlines in southern Iberia (Spain) during the Holocene. Quaternary Science Reviews 15, 773–780. Graham, J., 1988. Collection and analysis of field data. In: Tucker, M. (Ed.), Techniques in Sedimentology. Blackwell Scientific Publications, Oxford, pp. 5–62. Granja, H.M., De Groot, T.A.M., 1996. Sea-level rise and neotectonism in a Holocene Coastal Environment at Cortegac- a Beach (NW Portugal): a case study. Journal of Coastal Research 12 (1), 160–170. Harrison, S.P., Digerfeldt, G., 1993. European lakes as palaeohydrological and palaeoclimatic indicators. Quaternary Science Reviews 12, 233–248. Haslett, S.K., Strawbridge, F., Martin, N.A., Davies, C.F.C., 2001. Vertical saltmarsh accretion and its relationship to sea-level in the Severn Estuary, UK: an investigation using foraminefera as tidal indicators. Estuarine, Coastal and Shelf Science 52, 143–153.

Heiri, O., Lotter, A.F., Lemcke, G., 2001. Loss on ignition as a method for estimating organic and carbonate content in sediments: reproducibility and comparability of results. Journal of Paleolimnology 25, 101–110. Konert, M., Vandenberghe, J., 1997. Comparison of laser grain size analysis with pipette and sieve analysis: a solution for the underestimation of the clay fraction. Sedimentology 44, 523–535. Lentacker, A., 1986. Preliminary results of the fauna of ‘‘Cabeco de Amoreira’’ and ‘‘Cabeco de Arruda’’ (Muge, Portugal). Trabalhos de Antropologia e Etnologia 24, 9–26. Lobo, F.J., Herna´ndez-Molina, F.J., Somoza, L., Dı´ az del Rı´ o, V., 2001. The sedimentary record of the post-glacial transgression on the Gulf of Cadiz continental shelf (Southwest Spain). Marine Geology 178, 171–195. Long, A., 2001. Mid-Holocene sea-level change and coastal evolution. Progress in Physical Geography 25 (3), 399–408. Lubell, D., Jackes, M., Schwarcz, H., Knyf, M., Meiklejohn, C., 1994. The Mesolithic–Neolithic transition in Portugal: isotopic and dental evidence of diet. Journal of Archaeological Science 21, 201–206. Lucas, M.M., Ferrari, M., 1993. Relato´rios dos trabalhos realizados na a´rea da Pre´-Histo´ria da regia˜o. O Foral, Vol. 1, pp. 8–23. Mateus, J., Queiroz, P., 2000. Lakelets, lagoons, and peat-mires in the coastal plane South of Lisbon-Palaeoecology of the Northern Littoral of Alentejo. In: Excursion Guide of the Second Workshop of the Southern European Working Group of the European Lake Drilling Program (ELPD-ESF), Sintra, Portugal, pp. 33–37. Mateus, J.M., 1985. The coastal lagoon region near Carvalhal during the Holocene; some geomorphological aspects. Actas Reunia˜o do Quaterna´rio Ibe´rico 2, 237–249. Mendes Correa, A.A., 1933. Les migrations pre´historiques. Revue anthropologique 7–9, 1–26. Miall, A.D., 1996. The Geology of Fluvial Deposits. Sedimentary Facies, Basin Analysis and Petroleum Geology. Springer, New York, 582pp. Monges Soares, A.M., 1993. Isotope Techniques in the Study of Past and Current Environmental Changes in the Hydrosphere and the Atmosphere (Proceedings). International Atomic Energy Agency-SM329/49, Vienna, pp. 471–485. Moore, P.D., Webb, J.A., Collinson, M.E., 1989. Pollen analysis. Blackwell publishers, Oxford. Morais Arnaud, J.E., 1989. The mesolithic communities of the Sado Valley, Portugal, in their ecological setting. In: Bonsall, C. (Ed.), The Mesolithic in Europe. John Donald Publishers, Edinburgh, pp. 614–631. Morales, J.A., 1997. Evolution and facies architecture of the mesotidal Guadiana River delta (SW Spain-Portugal). Marine Geology 138, 127–148. Morhange, C., Vella, C., Provansal, M., Hesnard, A., Laborel, J., 1999. Human impact and natural characteristics of the ancient ports of Marseille and Fos in Provence, Southern France. In: Leveau, P., Trement, F., Walsh, F., Barker, G. (Eds.), Environmental Reconstruction in Mediterranean Landscape Archaeology. Oxbow Books, Oxford, pp. 145–153. Mozzi, P., Azevedo, M.T., Nunes, E., Raposo, L., 2000. Middle terrace deposits of the Tagus River in Alpiarc- a, Portugal, in relation to early human occupation. Quaternary Research 54, 359–371. Pape, T., 1996. Sample preparation for grain-size determination by laser diffraction. In: Buurman, P., Van Lagen, B., Velthorst, E.J. (Eds.), Manual for Soil and Water Analysis. Backhuys Publishers, Leiden, pp. 251–278. Passmore, D.G., Stevenson, A.C., 1999. Geoarchaeology of the lower Muge valley, west-central Portugal. Universities of Durham and Newcastle upon Tyne Archaeological Reports 1998, pp. 12–16. Peacock, J.D., 1993. Late Quaternary marine mollusca as palaeoenvirironmental proxies: a compilation and assessment of basic numerical data for Atlantic species found in shallow waters. Quaternary Science Reviews 12, 263–275.

ARTICLE IN PRESS T. van der Schriek et al. / Quaternary International 189 (2008) 135–151 Pilcher, J.R., 1991. Radiocarbon Dating. In: Smart, P.L., Francis, P.D. (Eds.), Quaternary Dating Methods—A User’s Guide. Technical Guide No. 4. Quaternary Research Organisation, London, pp. 17–26. Psuty, N.P., Moreira, E.S.A., 2000. Holocene sedimentation and sea level rise in the Sado Estuary, Portugal. Journal of Coastal Research 16 (1), 125–138. Reed, J.M., Stevenson, A.C., Juggins, S., 2001. A multi-proxy record of Holocene climatic change in Southwestern Spain: the Laguna de Medina, Ca´diz. The Holocene 11 (6), 707–719. Richards, M.P., Hedges, R.E.M., 1999. Stable isotope evidence for similarities in the types of marine foods used by Late Mesolithic humans at sites along the Atlantic coast of Europe. Journal of Archaeological Science 26, 717–722. Roche, J., 1977. Les amas coquilliers me´solithiques de Muge (Portugal). Chronologie, milieu naturel et leurs incidences sur le peuplement humain. Approche e´cologique del’homme fossile. Supple´mant au Bulletin de l’Association Franc- aise pour l’E´tude du Quaternaire 47, 353–359. Rodrı´ guez-Ramı´ rez, A., Rodrı´ guez-Vidal, J., Ca´ceres, L., Clemente, L., Belluomoni, G., Manfra, L., Improta, S., Ramo´n de Andre´s, J., 1996. Recent coastal evolution of the Don˜ana national park (SW Spain). Quaternary Science Reviews 15, 803–809. Rola˜o, J.M.F., 1994. Uma Nova Abordagem da Necro´pole da Moita do Sebastia˜o. Anais Serie Histo´ria (Universidade Auto´noma de Lisboa) 1, 9–16. Santos, L., Sa´nchez Gon˜i, M.F., 2003. Lateglacial and Holocene environmental changes in Portuguese coastal lagoons 3: vegetation history of the Santo Andre´ coastal area. The Holocene 13 (3), 459–464. Scott, D.B., Medioli, F.S., 1978. Vertical zonations of marsh foraminifera as accurate indicators of former sea-levels. Nature 272, 528–531. Soter, S., Blackwelder, P., Tziavos, C., Katsonopoulou, D., Hood, T., Alvarez-Zarikian, C., 2001. Environmental analysis of cores from the Helike Delta, Gulf of Corinth, Greece. Journal of Coastal Research 17 (1), 95–106. Stevenson, A.C., Battarbee, R.W., 1991. Palaeoecological and documentary records of recent environmental change in Garaet el Ichkeul: a seasonally saline lake in NW Tunisia. Biological Conservation 58 (3), 275–295. Stevenson, A.C., Harrison, R.J., 1992. Ancient forests in Spain: a model for land-use and dry forest management in south-west Spain from 4000 BC to 1900 AD. Proceedings—Prehistoric Society 58, 227–247. Straus, L.G., Altuna, J., Vierra, B., 1990. The Concheiro at Vidigal: a contribution to the Late Mesolithic of Southern Portugal. In: Vermeersch, P.M., Van Peer, P. (Eds.), Contributions to the Mesolithic in Europe. Leuven University Press, Leuven, pp. 463–474. Stuiver, M., Reimer, P.J., Bard, E., Beck, W.E., Burr, G.S., Hughen, K.A., Kromer, B., McCormac, F.G., Van de Plicht, J., Spurk, M., 1998a. INTCAL98 radiocarbon age calibration 0–24,000 BP. Radiocarbon 40, 1041–1083. Stuiver, M., Reimer, P.J., Braziunas, T.F., 1998b. High-precision radiocarbon age calibration for terrestrial and marine samples. Radiocarbon 40, 1127–1151.

151

Taylor, D.M., Pedley, H.M., Davies, P., Wright, M.W., 1998. Pollen and mollusc records for environmental change in central Spain during the mid- and late Holocene. The Holocene 8 (5), 605–612. Tucker Abbott, R., 1990. Seashells of the Northern Hemisphere. Dragon’s World Ltd., Surrey. USDA, 1975. Soil taxonomy: A Basic System of Soil Classification for Making and Interpreting Soil Surveys. United States Soil Conservation Service, Washington, DC. van der Knaap, W.O., van Leeuwen, J.F.N., 1994. Holocene vegetation, human imact and climatic change in the Serra da Estrela, Portugal. Dissertaciones Botanicae 234, 497–535. van der Knaap, W.O., van Leeuwen, J.F.N., 1995. Holocene vegetation succession and degradation as responses to climatic change and human activity in the Sierra de Estrela, Portugal. Review of Palaeobotany and Palynology 89, 153–211. van der Knaap, W.O., van Leeuwen, J.F.N., 1997. Late Glacial and early Holocene vegetation succession, altitudinal vegetation zonation, and climatic change in the Serra da Estrela, Portugal. Review of Palaeobotany and Palynology 97, 239–285. van Leeuwaarden, W., Janssen, C.R., 1985. A preliminary palynological study of peat deposits near an Oppidum in the Lower Tagus Valley, Portugal. Actas Reunia˜o do Quaterna´rio Ibe´rico 2, 226–236. Veiga Ferreira, O.d., 1954. Faune malacologique: crustaces et poissonsMuge (Moita do Sebastia˜o). Comunicac- oes dos Servic- os Geolo´gicos de Portugal 35, 339–352. Vierra, B.J., 1995. Portuguese Mesolithic research, subsistence and stone tool technology: an old world perspective. Anthropological Research Papers, vol. 47, Arizona State University, pp. 1–19. Vilar, H.V., 1993. Em torno do foral de Salvaterra de Magos: a genese dum concelho. O Foral, Vol. 1, 34–38. Yll, E.I., Pe´rez-Obiol, R., Pantaleon-Cano, J., Roure, J.M., 1999. Palynological evidence for climatic change and human activity during the Holocene on Minorca (Balearic Islands). Quaternary Research 48, 339–347. Zazo, C., Goy, J.-L., Somoza, L., Dabrio, C.-J., Belluomini, G., Improta, S., Lario, J., Bardaji, T., Silva, P.-G., 1994. Holocene sequence of sealevel fluctuations in relation to climatic trends in the Atlantic–Mediterranean linkage coast. Journal of Coastal Research 10 (4), 933–945. Zazo, C., Goy, J.L., Lario, J., Silva, P.G., 1996. Littoral zone and rapid climatic changes during the last 20,000 years. the Iberian study case. Zeitschrift fu¨r Geomorphologie, Suppl.-Bd. 102, 119–134. Zazo, C., Dabrio, C.J., Borja, F., Goy, J.L., Lezine, A.M., Lario, J., Polo, M.D., Hoyos, M., Boersma, J.R., 1999. Pleistocene and Holocene aeolian facies along the Huelva coast (southern Spain): climatic and neotectonic implications. Geologie en Mijnbouw 77 (3–4), 209–224. Zbyszewski, G., 1946. Etude Geologique de la region d’Alpiarc- a. Comunicac- o˜es dos Servic- os Geolo´gicos de Portugal 27, 145–268. Zilha˜o, J., 1993. The spread of Agro-pastoral economies across Mediterranean Europe: a view from the Far West. Journal of Mediterranean Archaeology 6 (1), 5–63. Zilha˜o, J., 2000. From the Mesolithic to the Neolithic in the Iberian peninsula. In: Price, T.D. (Ed.), Europe’s First Farmers. Cambridge University Press, Cambridge, pp. 144–182.