Improving our knowledge of rapid geomagnetic field intensity changes ...

Earth and Planetary Science Letters 355–356 (2012) 131–143

Contents lists available at SciVerse ScienceDirect

Earth and Planetary Science Letters journal homepage: www.elsevier.com/locate/epsl

Improving our knowledge of rapid geomagnetic field intensity changes observed in Europe between 200 and 1400 AD Miriam Go´mez-Paccard a,n, Annick Chauvin b, Philippe Lanos b,c, Philippe Dufresne b,c, Mary Kovacheva d, Mimi J. Hill e, Elisabet Beamud f, Sophie Blain c, Armel Bouvier c, Pierre Guibert c, Archaeological Working Team1 Institut de Cie ncies de la Terra Jaume Almera, ICTJA-CSIC, Sole´ i Sabarı´s s/n, 08028 Barcelona, Spain Geósciences-Rennes, UMR 6118, Universite´ de Rennes 1,CNRS, Campus de Beaulieu, Bˆ at. 15, CS 74205, 35042 Rennes, France c Centre de Recherches en Physique Appliqueé a l’Archeólogie, UMR 5060, CNRS, Universite´ de Bordeaux 3, France d NIGGG, BAS, Academician Georgi Bonchev Street, Block 3, 1113 Sofia, Bulgaria e Geomagnetism Laboratory, School of Environmental Sciences, University of Liverpool, Liverpool L69 7ZE, UK f Laboratori de Paleomagnetisme CCiT UB-CSIC, Institut de Cie ncies de la Terra Jaume Almera, Sole´ i Sabarı´s s/n, 08028 Barcelona, Spain a

b

a r t i c l e i n f o

abstract

Article history: Accepted 22 August 2012 Editor: P. DeMenocal

Available archaeomagnetic data indicate that during the past 2500 yr there have been periods of rapid geomagnetic field intensity fluctuations interspersed with periods of almost constant field strength. Despite Europe being the most widely covered region in terms of archaeomagnetic data the occurrence and the behaviour of these rapid geomagnetic field intensity changes is under discussion and the challenge now is to precisely describe them. The aim of this study is to obtain an improved description of the sharp intensity change that took place in western Europe around 800 AD as well as to investigate if this peak is observed at the continental scale. For this purpose 13 precisely dated early medieval Spanish pottery fragments, four archaeological French kilns and three collections of bricks used for the construction of different French historical buildings with ages ranging between 335 and 1260 AD have been studied. Classical Thellier experiments performed on 164 specimens, and including anisotropy of thermoremanent magnetisation and cooling rate corrections, gave 119 reliable results. The 10 new high-quality mean archaeointensities obtained confirm the existence of an intensity maximum of 85 mT (at the latitude of Paris) centred at 800 AD and suggest that a previous abrupt intensity change occurred around 600 AD. Together with previously published data from western Europe that we deem to be the most reliable, the new data also suggest the existence of two other abrupt geomagnetic field intensity variations during the 12th century and around the second half of the 13th century AD. High-quality archaeointensities available from eastern Europe indicate that very similar geomagnetic field intensity changes occurred in this region. European data indicate that very rapid intensity changes (of at least 20 mT/century) took place in the recent history of the Earth’s magnetic field. The results call for additional high-quality archaeointensities obtained from precisely dated samples and for a selection of the previously published data if a refined description of geomagnetic field intensity changes at regional scales is to be obtained. & 2012 Elsevier B.V. All rights reserved.

Keywords: archaeomagnetism archaeointensity secular variation dipole moment Thellier Europe

1. Introduction

n

Corresponding author. Tel.: þ34 934095410; fax: þ 34 934110012. E-mail address: [email protected] (M. Go´mez-Paccard). 1 ˜ avate (A´rea de Arqueologıá, Facultad Sonia Gutie´rrez-Lloret and Victor Can de Filosofıá y Letras, Universidad de Alicante, Spain), Christine Oberlin (Centre de datation par le RadioCarbone, Universite´ de Lyon 1, France), Christian Sapin (Laboratoire Artehis, CNRS, UMR 5594, Universite´ de Bourgogne, Dijon, France), Daniel Prigent (Service De´partemental de lArcheólogie, CG49, Angers, France), Se´bastian Jesset (Service archeólogique municipal, Orleáns, France), Herve´ Pomare des (INRAP: Institut national de recherche Archeólogique Pre´ventive, Direction interre´gionale Me´diterraneé, Nˆımes, France). 0012-821X/$ - see front matter & 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.epsl.2012.08.037

Unravelling past changes in geomagnetic field intensity is crucial to understand the dynamics of the geodynamo and to investigate the link between solar activity, 14C production, the Earth’s magnetic field and climate (e.g., Bard and Delaygue, 2008; Courtillot et al., 2007; Gallet et al., 2005; Muscheler et al., 2007; Snowball and Muscheler, 2007; Wollin et al., 1978). Geomagnetic field reconstructions can be obtained by studying well-dated archaeological material, volcanic rocks or lacustrine and marine sediments. Recently, much effort has been made to recover palaeosecular variation (PSV) of the geomagnetic field in Europe

132

´mez-Paccard et al. / Earth and Planetary Science Letters 355–356 (2012) 131–143 M. Go

over the last millennia (e.g., Genevey et al., 2009; Go´mez-Paccard et al., 2006a, 2008; Herve´ et al., 2011; Schnepp et al., 2009; Ma´rton, 2010). However, higher resolution intensity records are undoubtedly necessary to improve reconstructions of past geomagnetic field intensity changes (e.g., Go´mez-Paccard et al., 2008; Snowball and Muscheler, 2007; Tema et al., 2012). The need for a greater number of globally distributed palaeointensity data is also highlighted by the inconsistencies observed between global geomagnetic field models and data even for recent times (Finlay, 2008; Genevey et al., 2009; Gubbins et al., 2006; Hartmann et al., 2011; Jackson et al., 2000; Suttie et al., 2011). Another important consideration for geomagnetic field reconstructions is the reliability of the palaeointensity data (Chauvin et al., 2000). In this sense objective selection criteria based on palaeomagnetic considerations are clearly needed in order to assess the reliability of palaeointensities (Chauvin et al., 2000; Genevey et al., 2008; Go´mez-Paccard et al., 2008; Schnepp et al., 2009). On the other hand, the ages uncertainties (from few years when historical records are available up to some centuries) and errors associated with intensity determinations indicate that several data per century are necessary if a precise description of geomagnetic field intensity changes wants to be obtained (Go´mez-Paccard et al., 2006c). This multi-site approach has been widely applied for the construction of PSV directional curves (e.g., Gallet et al., 2002; Go´mez-Paccard et al., 2006b; Schnepp and Lanos, 2005; Tema et al., 2006; Zananiri et al., 2007). Available western European data indicate that during the past 2500 yr there have been periods of rapid intensity fluctuations, such as the one observed around 800 AD, interspersed with periods of little change, as seen during Roman times (Chauvin et al., 2000; Gallet et al., 2003, 2009a; Genevey and Gallet, 2002; Genevey et al., 2009; Go´mez-Paccard et al., 2008). The challenge now is to precisely describe these rapid intensity changes and to understand how such behaviour can arise (Gallet et al., 2009b). One of the periods with a strikingly poor data coverage is the period between 500 and 1000 AD for which only six archaeointensities are available for western Europe (Catanzariti et al., 2012; Gallet et al., 2009a; Genevey and Gallet, 2002; Kovacheva et al., 2009a). Based on the available French data an intensity peak around 800 AD followed by a strong decrease of about 30% with associated rate of change as fast as 7 mT/century has been previously identified (Genevey and Gallet, 2002). In order to further constrain field strength behaviour around 800 AD, new archaeointensities have been obtained from the study of 13 precisely dated early medieval Spanish pottery fragments with ages ranging from 750 to 900 AD. In addition, four French kilns and three collections of bricks used in the construction of historical French buildings with ages ranging between 335 and 1260 AD have been studied. The new data, together with a selection of previous archaeointensities, enable an improved description of geomagnetic field intensity changes that took place in western Europe between the 3rd and the 14th centuries AD. Finally, the selected western European data are compared with high-quality data from eastern Europe and with geomagnetic field models’ results.

2. Archaeological setting and dating 2.1. Early medieval Spanish pottery fragments Thirteen pottery fragments collected from El Tolmo de Minateda archaeological site (TM in Fig. 1) have been studied. The archaeological/historical context together with numismatic evidence and stratigraphic constraints between the different stratigraphic units where the material was found allows the

Fig. 1. Location of the archaeological/historical sites where the archaeomagnetic material was recovered. The reference location (Paris) where the archaeointensity data have been relocated is also shown.

precise dating of the studied ceramic fragments (see Table 1 and Supplementary material: Section 1 and Table S1, for further details). 2.2. Bricks used for the construction of historical French buildings The early medieval period in France (between 400 and 1100 AD, approximately) is not well known mainly due to the restricted number of buildings from this period still standing. Indeed most of the pre-Romanesque buildings have been reused, abandoned and savaged. Only structures constructed using nonperishable materials (e.g., stones or bricks), such as religious constructions, remained standing. Three historical buildings from western France (Fig. 1) have been studied from this period: the collegial church St. Martin located in Angers (site 49007I), the Notre-Dame-Sous-Terre church (site 50353) located in the MontSaint-Michel abbey and the keep of the castle of Avranches (site 50025A). In order to investigate the chronology of these sites a multidisciplinary approach was followed, involving experts of different dating techniques (thermoluminescence TL and radiocarbon 14C), art historians and archaeologists. TL was used to date the last heating of the bricks used for the construction of the historical buildings, which is in fact the date of the manufacture of the bricks. TL dating and archaeointensity analyses were performed on subsamples from the same unoriented cores, which were drilled directly from different bricks. Radiocarbon analyses were conducted on charcoals found within mortars. The final dates ascribed for the French buildings (Table 1) were obtained using a Bayesian statistical approach for chronological modelling (Lanos, 2004). This approach takes into account all the chronological information available for each site (archaeological and historical constraints and 14C and TL dating results) and assumes that the dates given by the different methods are related to the same event (i.e. the last firing of the set of bricks) with some errors that are modelled with a Gaussian prior density with unknown variance (Lanos, 2001, 2004; RenDateModel: http://www.projetplume.org/fr/relier/rendatemodel; Sapin et al., 2008). A Monte Carlo Markov Chain computation (Gilks et al., 1996) is performed over a large range of variances [0, þN], yielding to a posterior


133

Table 1 Location and description of the archaeological studied sites. The type of material collected as well as the main information on which the absolute dates proposed are based is indicated (see Supplementary material for further details). The directional archaeomagnetic results obtained for the four French kilns are given relocated to Paris (latitude 48.91N, longitude 2.31E). Dec and Inc, mean declination and inclination; n, number of samples used for calculation of the mean archaeomagnetic directions; k and a95, precision parameter and 95% confidence level of mean archaeomagnetic directions. Archaeological site (material)

Long.

(deg.)

(deg.)

Clermont- lHe´rault (kiln)

43.63N

3.43E

34079C

Archaeological context: end of Roman period [300, 600] þone 14C from neighbouring kiln

Dec¼ 0.41, Inc ¼62.51 n ¼20, a95 ¼ 1.11, k¼ 875

[335, 540]

Pierrefitte-sur-Sauldre (kiln)

47.52N

2.10E

41176-9

Archaeological context: local ceramics [600, 650] þ one 14C date for the kiln

Dec¼ 6.11, Inc¼ 70.21 n ¼21, a95 ¼ 1.11, k ¼791

[590, 655]

Neóns-sur-Creuse (kiln)

46.74N

0.91E

36137A

Archaeological context: Early Middle Age [600, 900]

Dec¼ 5.31, Inc¼ 72.31 n ¼14, a95 ¼ 1.61, k ¼550

[600, 900]

Saran, Zac des Vergers (kiln)

47.95N

1.88E

45302A

Archaeological context: local ceramics [675, 725]

Dec¼ 3.01, Inc ¼73.71 n ¼13, a95 ¼ 1.91, k¼ 412

[675, 725]

Angers, Collegial church St. Martin (bricks in buildings)

47.47N

0.55W

49007I

Five thermoluminescence results þ one 14 C date

[810, 910]

Mont-Saint-Michel, NotreDame-Sous-Terre church (bricks in buildings)

48.63N

1.5W

50353

Eight thermoluminescence results þthree 14 C dates

[900, 990]

Avranches, Castle of Avranches 48.68N (bricks in buildings)

1.35W

50025A

Five thermoluminescence results

[1100, 1260]

Albacete, Tolmo de Minateda (ceramic fragments)

1.61W

TM

Historical/archaeological constraints

Attributed ages (yrs AD) [800, 850] [750, 800] [775, 800] [750, 800] [750, 800] [750, 800] [750, 800] [750, 800] [750, 800] [750, 800] [825, 875] [775, 800] [850, 900]

38.48N

Lab. code

Dating method

Archaeomagnetic direction (relocated to Paris)

Dating at 95% confidence level

Lat.

(yr. AD)

TM-02 TM-03 TM-04 TM-05 TM-06 TM-07 TM-08 TM-09 TM-10 TM-11 TM-12 TM-13 TM-14

dating density from which an interval at a given confidence level can be determined. In this study we used the 95% confidence level. The final dates ascribed for each site are described in Table 1 (see Supplementary material: Section 2 and Table S2 for further details). 2.3. French archaeological kilns Four mediaeval kilns were sampled at Clermont-l’He´rault (site 34079C), Pierrefitte-sur-Sauldre (site 41176-9), Neón-sur-Creuse (36137A) and Saran (45302A) (Fig. 1). At Neóns-sur-Creuse a hearth of a 1 m diameter kiln was excavated, while larger pottery kilns (up to 4 m length on 2 m wide) were recovered in Saran, Pierrefitte-sur-Sauldre and Clemont-l’He´rault. Kilns 36137A and 45302 were dated by archaeological constraints (Supplementary material: Section 3 for further details). Two 14C results obtained from charcoal samples found during the excavations of sites 34079C and 41176-9 were also obtained. The Bayesian approach was used to date the last firing of kilns 34079C and 41176-9 by combination of the archaeological information and the calibrated 14 C results (Supplementary material: Section 3). Additionally, mean directions of the characteristic remanent magnetisation (ChRM) of each kiln were derived from thermal

demagnetisation experiments performed on one specimen per core and from principal component analysis (Kirschvink, 1980) and Fisher statistics (Fisher, 1953). Experiments were performed at the Geósciences-Rennes laboratory and were corrected for the effect of the thermoremanent magnetisation (TRM) anisotropy. Only one component of magnetisation was observed on samples from sites 34079C, 36137A and 41176-9, corresponding to median destructive temperatures (MDT) between 300 and 510 1C. Precise (a95 o21) mean site directions for kilns 34079C, 36137A and 41176-9 were calculated using 20, 14 and 21 samples, respectively (Table 1). The mean directions were relocated to Paris using the Virtual Geomagnetic Pole (VGP). A secondary component of magnetisation removed at temperatures between 300 and 460 1C was observed on 14 samples from kiln 45302A. On these samples, the high temperature component has no preferential direction in in situ coordinates suggesting that it corresponds to the heating of the tiles (see Section 3.1) during their manufacture. The other 13 samples (corresponding to MDT between 300 and 450 1C) show a well-defined single component of magnetisation and were used to derive mean direction for kiln 45302A. The ChRM directions obtained (Table 1) are in close agreement with the directions described by the French PSV curve (Gallet et al., 2002) for the dates proposed for the abandonment of

134


each kiln (Table 1). The directional analysis gives further support to the chronological framework proposed. However, it should be noted that archaeomagnetic dating has not been used to refine the final ages ascribed for each kiln.

3. Methodology 3.1. Sample preparation Four specimens of about 1 cm 1 cm per pottery fragment were prepared for the Thellier experiments. The 52 specimens obtained were packed into salt pellets in order to treat them as standard palaeomagnetic cubes (Pick and Tauxe, 1993). The remaining material was kept for microwave and rock magnetic experiments. Two mini cores per fragment (5 mm diameter and 2–5 mm long) were drilled for use with the 14 GHz microwave and Tristan magnetometer system (Stark et al., 2010) at Liverpool University. Standard palaeomagnetic cores were drilled from bricks in the Notre-Dame-Sous-Terre church (90 cores), the Collegial church St. Martin (166 cores) and the castle of Avranches (10 cores). Large oriented block samples were taken from tiles (sites 45302A, 41176-9, 36137A) and bricks (sites 34079C) making up the sides of the four studied medieval kilns. The blocks were oriented by placing plaster hats on the top surface (levelled horizontal using a bubble) and oriented using sun and magnetic compasses. A circular saw was used to remove the block samples that were drilled at the laboratory giving standard oriented palaeomagnetic cores from each block sample.

procedure, which includes four additional heating–cooling steps (Go´mez-Paccard et al., 2006c). Several criteria were used to select acceptable Thellier experiments at the specimen level. pTRM checks have been considered positive if, at a given temperature, the difference between the original pTRM and the pTRM check does not exceed 10% of the total TRM acquired. We fixed a limit of 45% for the fraction of the initial natural remanent magnetisation (NRM) involved for archaeointensity determinations (f parameter; Coe et al., 1978). Only linear NRM–TRM diagrams corresponding to well defined straight lines going to the origin in the Zijderveld diagrams have been considered. The maximum angular deviation (MAD; Kirschvink, 1980) and the deviation angle (DANG; Pick and Tauxe, 1993) have to be lower than 51. However, it should be noted that the DANG is not always sufficient criteria to detect mineralogical changes or multidomain behaviour if the direction of the NRM and the magnetic laboratory field are subparallel during Thellier experiments (Herve´ et al., 2011). Microwave experiments were also conducted on selected ceramic fragments as an additional check on repeatability and reliability of the more extensive Thellier experiments. One specimen per fragment underwent microwave demagnetisation in order to determine remanence characteristics and the powers needed to demagnetise the samples. The sister specimens then underwent the Coe variant of the Thellier method (Coe, 1967) incorporating microwave pTRM (pTMRM) and pTMRM tail checks generally every second step. Additional pTMRM checks back to the lowest microwave power were also carried out at various higher powers. The applied field of 70 or 75 mT was imparted along the NRM direction in order to minimise anisotropy and multidomain effects (e.g., Biggin, 2010).

3.2. Magnetic measurements 4. Rock magnetism 4.1. Q ratios Initial NRM intensities and susceptibilities have been measured for all the specimens (Fig. 2; Supplementary material: Table S3). The Koenigsberger ratios Q (defined as the ratio of NRM to the induced magnetisation using a field of 40 A/m) range between 1 and 1000. Some of the material analysed (some TM ceramic fragments) acquired a weak partial TRM with Q values lower than

100000

NRM intensity (mA/m)

Classical Thellier experiments were conducted at the palaeomagnetic laboratories of the Institute of Earth Sciences Jaume Almera (Barcelona, Spain) and Geósciences-Rennes (Rennes, France). The thermal treatment was conducted using MMTD-80 (Magnetic Measurements) furnaces and the magnetisation measurements were performed using a SRM755R (2G Enterprises) three-axes cryogenic superconducting rock, Agico JR5 and Molspin spinner magnetometers. Bulk susceptibilities (w) were also measured with a Kappabridge KLY-2 (Geofyzica Brno) susceptibility bridge using a field of 0.1 mT at a frequency of 470 Hz and with a Bartington susceptibility metre. The original Thellier method (Thellier and Thellier, 1959) with regular partial thermoremanent magnetisation (pTRM) checks was used to estimate archaeointensities. Experiments were carried out in air and a laboratory field of 40, 50 or 60 mT was applied. At each temperature step, the samples were first heated and cooled with the laboratory field applied along their Z-axis and second, were heated and cooled with the laboratory field applied in the opposite sense. The experimental procedures followed are the same as in our previous work (Go´mez-Paccard et al., 2006c, 2008). Biases due to TRM anisotropy in some archaeological material such as bricks, tiles or ceramics should be taken into account if accurate archaeointensities are to be obtained (Chauvin et al., 2000; Genevey et al., 2008). Therefore the TRM anisotropy tensor was determined for each specimen and used to correct archaeointensities. Another important consideration in archaeointensity studies is the cooling rate dependence of TRM acquisition. Although the real cooling time experienced by the different pottery fragments, kilns or bricks is not exactly known, a slow cooling time of about 24 h was used during the cooling rate protocol carried out in the laboratory. A cooling rate correction factor was then obtained for each specimen if no major evolution of TRM acquisition capacity was observed during the entire

10000

1000

100

10

50353 34079C 49007I 36137A 50025A 41176-9 45302A TM

Q=

100

Q=

0

100

Q=

10

Q=

0.1

1

11

0

Susceptibility (10-3 SI) Fig. 2. Intensity of the natural remanent magnetisation (NRM) versus bulk susceptibility. For the specimens derived from the ceramic fragments recovered at El Tolmo de Minateda (TM) these parameters have been normalised by the weight. Black lines indicate constant Koenigsberger ratios (Q).


135

10. The other specimens (from kilns, bricks and the remaining TM samples) show Q between 10 and 1000. This indicates a strong TRM acquired during the manufacture of the bricks and ceramics or during the last use of the kilns.

phase during heating at around 560 1C with a Curie temperature of 660 1C. On cooling this phase is no longer present, with the Curie temperature of the cooling curve around 580 1C (pure magnetite).

4.2. Spanish pottery fragments

4.3. French kilns and bricks used for the construction of historical French buildings

Rock magnetic experiments of ceramic fragments were carried out using a Variable Field Translation Balance at the University of Liverpool. Acquisition of isothermal remanent magnetisation (IRM), back-field coercivity, hysteresis loops and magnetisation versus temperature experiments were carried out on small pieces ( 100 mg) from each fragment. Samples TM-03, 05, 06, 07, 08, 09 and 11 are dominated by a low coercivity phase as evidenced by saturation of the IRM curves and closure of the hysteresis loops (Fig. 3a). The rest of the samples also contain a higher coercivity phase as evidenced by hysteresis loops that do not close and IRM curves that do not saturate (Fig. 3b). Wasp waisted hysteresis loops indicate the presence of these two phases. All samples (apart from TM-03) are dominated by a single Curie temperature ranging between 500 and 550 1C consistent with an impure magnetite phase. After heating to 700 1C all samples show a reduction in magnetisation on cooling, suggesting oxidation of magnetite or maghaemite to haematite. The Curie curve for sample TM-03 (not shown) indicates creation of a new magnetic

IRM acquisition curves were performed with an ASC pulse magnetiser, and alternating field (AF) demagnetisation was conducted with a Schonstedt GSD1. An Agico KLY3 and a CS-3 furnace were used for Curie experiments (K–T curves). Measurements of the remanent magnetisation were performed with an Agico JR5 magnetometer. All the experiments were carried out at the Geósciences-Rennes and Sofia palaeomagnetic laboratories. Samples of tiles and bricks from kilns are characterised by Curie temperatures (Tc) between 565 and 590 1C (Fig. 4a), i.e. very close to the Tc of magnetite. Only for two samples Tc’s close to 610 1C were observed (samples 41176-6 and 45302A-12), suggesting some highly oxidised magnetic mineral, like stable maghemite. K–T curves are most of the time reversible (Fig. 4a), in agreement with the thermal stability of magnetic minerals observed during the Thellier experiments (see below). Typical Median Destructive Fields (MDF) between 15 and 25 mT indicate low coercivity ferromagnetic grains. Saturation of the remanence

Sample TM-05

1

Sample TM-04

Normalised IRM

Normalised IRM

1

0

0 0

200

400

600

0

0.5

-500

500 -0.5

1000

Field (mT)

-1.0

Normalised magnetisation

1.0

-1000

1

0 0

200 400 600 Temperature (ºC)

200

400

600

Magnetic field (mT)




Magnetic field (mT)

1.0 0.5

-1000

-500

500 -0.5

1000

Field (mT)

-1.0 1

0 0

200 400 600 Temperature (ºC)

Fig. 3. Rock magnetic properties from representative ceramic fragments. Normalised IRM acquisition curves, hysteresis loops and normalised saturation magnetisation versus temperature curves are plotted. Dark (light) grey symbols on the thermomagnetic curves indicate the heating (cooling) branches. Sample TM-05 (a) was rejected for archaeointensity determinations and sample TM-04 (b) provided suitable archaeointensity results for two specimens.


Magnetic susceptibility (arbitrary units)

136

200 100

100

50025A-9

36137A-19 0

0 0

200

400

600ºC

0

200

Magnetic susceptibility (arbitrary units)

400

600ºC

Temperature (ºC)

Temperature (ºC)

400

60 50

300 40 200

30 20

100 10 49007I-248

50353-C32

0

0 0

200

400

600ºC

0

200

Temperature (ºC) 50353-A5

1

400

600ºC

Temperature (ºC) 50353-C32

49007I-248A

1B

50353-B

1

Normalised IRM

-A8 50353

49007I-32

Normalised IRM

Back-field

49007I-32 mT

10

60 50025A-10 36137A-17 50025A-7

-1

0

200

400

600

800

1000

1200

1400

1600

1800

Applied field (mT) Fig. 4. Rock magnetic properties from representative samples from French kilns and bricks used for the construction of historical French buildings. Magnetic susceptibility versus temperature curves (a), and isothermal remanent magnetisation (IRM) acquisition curves and back-field IRM curves (b).

is reached at between 200 and 250 mT while coercivity of remanence (Hcr) and saturated remanent magnetisation (Jrs) are close to 20 mT and 3.5 A/m, respectively (Fig. 4b). The presence of multidomain grains (MD) was tested on 16 samples, following the approach proposed by Shcherbakova et al. (2000). A pTRM was given between 300 1C and room temperature before to be demagnetised between the same temperature interval. The intensity of the pTRM tail is between 9% and 0.5% of the pTRM suggesting that the magnetic grains carrying the TRM are single domain (SD) or pseudo SD grains (Shcherbakova et al., 2000). Samples from bricks in buildings are characterised by Tc’s between 505 and 570 1C, with reversible K–T curves (Fig. 4a). MDF

ranges from 16 to 23 mT and Hcr between 18 and 55 mT. Strong values of Jrs are obtained, between 48 and 242 A/m. On some samples IRM saturation is not reached until 1000 mT (e.g., 50353A8 and 49007I-248A, 49007I-32 in Fig. 4b). However, these samples have Tc’s between 510 and 560 1C. All these characteristics indicate that the principal magnetic carriers are slightly substituted magnetites. Haematite was never observed. Despite no hysteresis measurement were performed for these samples, the high success rate and the linearity of the NRM–TRM diagrams over a large temperature interval during the Thellier experiments (see Section 5) suggest that magnetic grains carrying the TRM are SD grains.


Specimen: TM-04A f = 0.84 q = 40.2 F= 88.2 ± 1.7 μT

100 ºC

1

Specimen: TM-13A f = 0.80 q = 75.3 F= 76.4 ± 0.7 μT

100 ºC

1

137

Remaining NRM

Remaining NRM

300 NRM

250

+X

-Y

330 390

+Z

NRM +Y

+X

360 420

480 480

0

+Z

0 0

1

1

0

TRM gained

TRM gained

Specimen: 34079C-1P f = 0.93 q = 115.5 F= 73.7 ± 0.5 μT +X

Remaining NRM

360

1

Specimen: 36137A-19A f = 1.00 +Y q = 26.6 F= 79.2 ± 2.7 μT

100 ºC 240

Remaining NRM

100 ºC 1 240

+Y 510

560

+Z

330 -X

400

NRM

510

NRM

620

0

560

0 0

1

0

1

TRM gained

TRM gained

Specimen: 50025A-7B1 f = 0.88 q = 71.4 +X F= 62.4 ± 0.7. μT

100 ºC 350 400

1

Specimen: 490071-42B f = 0.89 q = 33.8 F= 82.1 ± 1.5 μT +Y

+Y

450 475

Remaining NRM

100 ºC

500

NRM

525 +Z

-X

Remaining NRM

1

+Z

480

NRM 200 250

0

555

0 0

+Z

400 450

550

1

0

1

TRM gained

TRM gained Specimen: 45302A-15P (rejected) 240

Specimen: TM-12 (rejected) 300 ºC

1.8

+Y

240 ºC

1 +X

Remaining NRM

-Y

470

+Z

330

Remaining NRM

NRM 390

NRM -X

150 100

+Z

400

560 540

0 0

1 TRM gained

2

0 0

4.4 TRM gained

Fig. 5. Thellier archaeointensity results. Typical NRM–TRM diagrams. Examples of accepted (a–f) and rejected (g and h) archaeointensity results of representative specimens corresponding to the three types of material analysed (Spanish ceramic fragments, samples taken from archaeological French kilns and bricks from historical French buildings). The NRM–TRM diagrams are shown together with the Zijderveld diagrams, in sample coordinates. In the Zijderveld diagrams open (solid) circles are projections upon vertical (horizontal) planes. NRM-TRM plots are normalised to the initial NRM intensity and closed circles correspond to data used for archaeointensity determinations. The estimated archaeointensity (F), the NRM fraction used (f) and the corresponding quality factor (q) are indicated.


138

5. Results: archaeointensity determination 5.1. Spanish pottery fragments 5.1.1. Thellier experiments The Thellier experiments were attempted on 52 specimens. The specimens are characterised by NRM intensities lower than 7 10 2 A/mg and susceptibilities lower than 10 3 SI units/g (Supplementary material: Table S3). Nine to fifteen steps were performed between 100 1C and the maximum temperature applied, which varies from 480 to 590 1C. Two types of behaviour have been observed. Specimens from samples TM-02, 03, 04, 06, 07, 09, 13 and 14 showed a well-defined straight line going towards the origin of the Zijderveld plots after removal of a very weak secondary component of magnetisation during the first temperature steps. This characteristic component likely corresponds to the TRM acquired during the manufacture of the pottery fragments. The maximum unblocking temperatures range between 480 and 540 1C, which is in agreement with the impure magnetite-like phase observed in the rock magnetic experiments. For these specimens linear NRM–TRM diagrams have been observed (Fig. 5a and b). Specimens from samples TM-05, 08, 10, 11 and 12 show a more complex behaviour, with two clear components of magnetisation in the Zijderveld diagrams (e.g., Fig. 5g). In order to only consider the most reliable archaeointensities we discarded the estimations based on these samples. The 23 accepted archaeointensity determinations which satisfied the selection criteria are listed in Supplementary material: Table S3. The NRM fraction used for archaeointensity determination ranges between 0.46 and 0.95 corresponding with quality factors q values between 13.6 and 75.3. No control of the exact position of the different specimens inside the salt pellets was possible. Therefore, no inferences about the direction of the principal axes of the TRM anisotropy tensor can be done. However, differences between uncorrected and TRM anisotropy corrected values can be very important in this kind of material (Chauvin et al., 2000) and must be taken into account. For the majority of the specimens, differences between the uncorrected and TRM anisotropy corrected values are lower than 30% (Fig. 6a). For specimens TM-06A, 06B, 03D and 14A differences between the uncorrected and corrected archaeointensities are higher (Supplementary material: Table S3). As in our previous work (Go´mez-Paccard et al., 2006c, 2008) archaeointensities have been corrected when the cooling rate correction factor obtained is

bigger than the estimated magnetic alteration that took place during the cooling protocol (Supplementary material: Table S3). In general, the cooling rate correction factors applied (Fig. 6b) are not very high with a maximum value of 7%. From the 52 specimens analysed only 23 archaeointensity determinations have been accepted, showing the difficulty to obtain reliable intensities in more than 50% of the analysed specimens. However, all the archaeointensities retained correspond to high-quality NRM–TRM diagrams and have been corrected for TRM anisotropy and cooling rate effects. In order to further investigate reliability and consistency of the results obtained, microwave experiments were carried out on selected samples.

5.1.2. Microwave experiments The scarcity of material from samples TM-03, 06 and 08 coupled with their low magnetisation meant that microwave experiments were not pursued for these fragments. Microwave demagnetisation on the remaining fragments demonstrates that a primary component of remanence can be isolated in all cases. A small secondary component of magnetisation is also present. Consistent with the thermal results, samples TM-11 and 12 have more complicated behaviour with an additional lower blocking spectra component of remanence. Sample TM-10, however, did not exhibit this behaviour in contrast to the Thellier results and the sister sample used for the archaeointensity experiment. Sample TM-05 revealed a large secondary component of magnetisation during demagnetisation. However, this was not present in the sister mini core used for the archaeointensity experiment. The secondary component of remanence, when present, does not seem to be homogenously distributed throughout the fragments. This could be explained by the large size (several cm) of some of the pottery fragments. However, it is worth noticing that this is generally not observed in Thellier results where very similar behaviour is observed in sister specimens analysed per fragment. Microwave archaeointensity results are summarised in Table S4 and successful examples shown in Fig. S1 of Supplementary material. Using the same acceptance criteria as for the Thellier experiments five of the eight fragments dominated by a single component of remanence were accepted. Sample TM-12 did not yield a successful result (failure of a pTRM check) consistent with the Thellier study. In contrast to the Thellier study which gave acceptable results for samples from fragment TM-09, pTRM checks failed during the microwave experiment. The microwave

30

40 35 Kilns Bricks Ceramic fragments

20

number of specimens

number of specimens

25

15 10 5

30 25 20 15 10 5 0

0 -20

0

20

40

60

TRM anisotropy correction factor (%)

80

100

-20

0

20

40

Cooling rate correction factor (%)

Fig. 6. TRM anisotropy and cooling rate correction factors. Differences between the uncorrected and the TRM corrected intensities (TRM anisotropy correction factor) at specimen level (a), and cooling rate correction factor applied to archaeointensity determinations (b). Each kind of material analysed is plotted separately.


experiment on sample TM-04 failed due to the experiment having to be abandoned before complete demagnetisation (low f factor). It was not possible to obtain archaeointensity estimates from samples TM-10 and 11 which contained significant secondary components of remanence. The microwave archaeointensities can be compared to the Thellier results. Cooling rate corrections have not been attempted, thus microwave results are compared to the TRM anisotropy corrected Thellier results. Microwave results from samples TM-07 and 14 are within the range obtained with the Thellier method (differences between the two methods are less than 4%) whereas the microwave result from TM-02 is 13% higher than the Thellier derived mean. A significant discrepancy is found for TM-13 with the microwave result nearly 30% higher than the mean of the Thellier estimates. There is no immediately obvious reason for this anomalously high result and further study would be necessary to see if the microwave result is repeatable. We do note, however, that TM-13 does yield the highest archaeointensity estimate in classical Thellier experiments. As only one experiment has been performed per fragment and more experiments are needed in order to confirm the microwave results, these data have not been used in order to recover past archeointensities. 5.2. French kilns and bricks used for the construction of historical French buildings Thellier experiments were attempted on 112 specimens, 96 of them gave reliable results. This high success rate shows the excellent physical and chemical stability of the magnetic minerals contained in the kilns and brick samples. This is in agreement with the magnetic results described in Section 4. All accepted results correspond to high-quality NRM–TRM diagrams related to very well-defined single components of magnetisation going through the origin (Fig. 5c–f). Maximum unblocking temperatures between 480 and 620 1C were observed. The rejected specimens showed a more complex behaviour with two components of magnetisation (e.g., Fig. 5h) and/or alteration of the magnetic mineralogy during the Thellier experiments. TRM anisotropy and cooling rate effects upon TRM intensity have been also investigated in these materials (Fig. 6; Supplementary material: Table S3; Kovacheva et al., 2009b). As expected, a low TRM anisotropy effect upon archaeointensities was found for the kilns, for which very low differences between uncorrected and TRM corrected intensities were generally obtained (Fig. 6a). This effect is more important for the bricks with differences reaching values up to 26%. The majority of the

139

specimens (from kilns or bricks) gave low cooling rate correction factors (less than 10%). However, in some cases these factors are bigger reaching values up to 25% (Fig. 6b).

6. Discussion 6.1. New archaeointensity data for western Europe Supplementary material: Table S3 summarises the results obtained at specimen level for the ceramic fragments. In general, archaeointensities derived from sister specimens from the same fragment are very consistent (e.g., TM-02A, 02B and 02C). However, some important differences are also observed (e.g., TM-03B and 03D), indicating that several specimens per fragment must be studied. Therefore, we retained only mean intensities calculated using at least two specimens per fragment. Additionally, maximum values of 10% around the means were accepted. Eight fragment-mean intensities satisfied these criteria (TM-02, 03, 04, 06, 07, 09, 13 and 14) and were calculated considering the weighting factor proposed by Pre´vot et al. (1985). Fragmentmean intensities are generally well defined with standard deviations of less than 8% for six of the eight fragments. In general, the results are very consistent for fragments which correspond to the same time period, but some differences are also observed (e.g., TM-06 and 13). Considering that the studied collection has a local character, we decided to group the fragments in three sets corresponding to the same age interval. The collection therefore comprises three groups with distinct ages. The different fragments of each group were considered together for deriving the group-mean intensities for each set. Mean intensities (relocated to Paris through the VADM) of 80.275.7 mT, 88.576.0 mT and 84.877.2 mT were obtained for ages ranging from 750 to 800 AD, from 775 to 800 AD and from 800 to 900 AD, respectively (Table 2). Supplementary material: Table S3 also shows the results obtained at specimen level for the French sites. Samples from the same site are generally very consistent (e.g., site 36137A). However, important differences are observed between some samples of the same site (e.g., 49007I-9BB and 249B). This highlights the need to study several samples per site in order to recover accurate mean intensities. Between four and 37 samples were used to determine the mean intensities. The means were calculated considering the weighting factor proposed by Pre´vot et al. (1985). Intensity values between 66.575.4 mT and 90.578.7 mT (relocated to the latitude of Paris) were obtained for ages ranging from 335 to 1260 AD (Table 2).

Table 2 New archaeointensity data for western Europe. Archaeological site; name of the archaeological site where the material was recovered; Lat. and Long., latitude and longitude of the site; Lab. code, laboratory code for each archaeomagnetic site or group; Age (yr AD), age of each site/group; N, number of independent samples or fragments retained per site/group; n, number of specimens retained per site/group; F 7 sd, site/group mean intensity and standard deviation before TRM anisotropy and cooling rate corrections; Fm 7sd, site/group mean intensity and standard deviation corrected for TRM anisotropy and cooling rate effects; Fpo, site/group weighted mean intensity; Fpa, site/group weighted mean intensity relocated to Paris; VADM, values of the virtual axial dipole moment. Archaeological Site

Lat. (deg.)

Clermont lHe´rault Pierrefitte-sur-Sauldre Neóns-sur-Creuse Saran, Zac des Vergers Albacete, Tolmo de Minateda

43.63 47.52 46.74 47.95 38.48

Albacete, Tolmo de Minateda Albacete, Tolmo de Minateda Angers, Collegial church St. Martin Mont-Sain- Michel, Notre-Dame-SousTerre church Avranches, Castle of Avranches

38.48 38.48 47.47 48.63

Long. (deg.)

Lab. code

Age (yr AD)

N N N N N

3.43 2.10 0.91 1.88 1.61

E E E E W

[335, [590, [600, [675, [750,

540] 655] 900] 725] 800]

N N N N

1.61 1.61 0.55 1.50

W W W W

34079C 41176-9 36137A 45302A TM-03þ TM-06 þ TM-07þ TM-09 TM-04 þTM-13 TM-02 þTM-14 49007I 50353

[775, [800, [810, [900,

800] 900] 910] 990]

50025A

[1100, 1260]

48.68 N

1.35 W

N

n

F7 sd (mT)

Fm7 sd (mT)

Fpo (mT)

Fpa (mT)

VADM (1022 Am2)

4 6 5 9 4

4 8 7 9 11

71.17 7.5 96.57 9.7 77.97 4.0 81.47 4.7 58.47 11.8

64.1 7 3.2 88.6 7 8.7 66.0 7 3.7 77.6 7 8.3 68.2 7 5.7

65.6 89.3 65.7 79.9 68.3

69.2 90.5 67.1 80.6 80.2

10.9 14.2 10.6 12.7 12.0

2 2 23 37

6 6 23 37

81.87 2.5 60.87 1.4 77.87 8.0 70.07 9.7

75.2 7 6.0 72.3 7 7.2 80.4 7 10.8 68.8 7 7.0

75.4 72.3 81.3 69.9

88.5 84.8 82.4 70.1

13.3 12.7 13.0 11.0

10

10

69.97 6.7

66.5 7 5.4

66.4

66.5

10.5


140

Ten new archaeointensities were obtained using one of the more reliable methods (Thellier and Thellier, 1959) including TRM anisotropy and cooling rate corrections. They were derived from the analysis of several specimens per pottery fragment and several fragments per age interval or several independent samples per structure/building. Consistent values were obtained from

the different materials studied: bricks, kilns and pottery fragments (Fig. 7a). For example, the results obtained for the 9th century from the ceramic fragments and bricks are very similar (84.8 mT and 82.4 mT, respectively). Therefore, we consider the new data as reliable markers of past geomagnetic field intensity in western Europe. As expected, the virtual axial dipole moment

western Europe, N=58 This study: kilns bricks ceramics

80

60

40 200

100 Archaeointensity (μT)

Archaeointensity (μT)

100

eastern Europe, N=42

Spain and Morocco France, Switzerland and Belgium

600

1000

80

60

Bulgaria Greece and Italy 40 200 600

1400

Age (AD)



80

60

Bayesian curve (this study)

600 1000 Age (AD)

60

Bayesian curve (this study)

600 1000 Age (AD)

1400

100



80

40 200

1400

100

80

60

40 200

1400

100

100

40 200

1000

Age (AD)

SCHA.DIF.3K ARCH3K.1

600 1000 Age (AD)

1400

80

60

40 200

SCHA.DIF.3K ARCH3K.1

600 1000 Age (AD)

1400

Fig. 7. Geomagnetic field intensity changes in Europe. New archaeointensities obtained in this study and selected previous results for western Europe and northern Morocco (a), and selected archaeointensities available from Bulgaria, Greece and Italy (b). In (c) and (d) the mean Bayesian curve derived from the selected data is shown in orange together with the 95% error envelope (green area). The SCHA.DIF.3K regional model (Pavoń-Carrasco et al., 2009) valid for Europe and the global model ARCH3K.1 (Korte et al., 2009) are shown in (e) and (f). Archaeointensities and models are referred to Paris (western Europe) and Sofia (eastern Europe). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)


(VADM) results obtained from the 8th to the 10th centuries (Table 2) are higher than the present-day dipole field value (Genevey and Gallet, 2002). 6.2. Geomagnetic field intensity changes in western Europe between 200 and 1400 AD In order to analyse geomagnetic field intensity changes in western Europe between 200 and 1400 AD previously published archaeointensities for Spain, France and nearby countries have been selected. As discussed by Chauvin et al. (2000) and Genevey et al. (2008) not all the archaeointensity data can be regarded as equally reliable. Therefore, we consider only high-quality archaeointensities obtained from Thellier or Thellier-derived experiments including pTRM checks during laboratory experiments. Only 48 archaeointensities from western Europe and nearby countries pass this selection criterion. The selection includes data from France (N ¼26, Chauvin et al., 2000; Gallet et al., 2009a; Genevey and Gallet, 2002; Genevey et al., 2009), Spain (N ¼ 18, Catanzariti et al., 2012; Go´mez-Paccard et al., 2006c, 2008), Switzerland (N ¼1, Donadini et al., 2008; Kovacheva et al., 2009a), Belgium (N ¼2, Genevey et al., 2009; Spassov et al., 2008) and northern Morocco (N ¼1, Go´mez-Paccard et al., 2012). All data have been relocated to the latitude of Paris through the VADM (Fig. 7a). The 48 selected data together with the new archaeointensities obtained in this study enable an improved description of geomagnetic intensity changes that took place in western Europe between the 3rd and the 14th centuries AD. The pattern of archaeointensities indicates approximately constant values between 300 and 500 AD (around 65 mT) followed by an increase up to 90 mT around 600 AD (new data of site 41176-9). However, the archaeointensity, recently obtained by Catanzariti et al. (2012) for similar ages, is not in agreement with this and gives a much lower value (yellow triangle in Fig. 7a). There are two possible explanations for the observed offset. The first is that the Earth’s magnetic field varied very quickly in a very short interval of time, and the second is that these sites are not accurately dated. With the available information we cannot consider one of these hypotheses to be more plausible than the other. Our new data (sites 41176-9, 45302A and 36137A, orange circles in Fig. 7) do however suggest that an important decrease of the Earth’s magnetic field intensity occurred between 600 and 750 AD which is also supported by previously published data (blue square, Fig. 7a). High intensities (of about 85 mT) are reached again very quickly around 800–850 AD. The occurrence of this maximum is now supported by seven archaeointensities (four obtained here and three previously published, Fig. 7a). A global trend of intensity decrease (better described now due to the new data of site 50353) is observed between 800 AD and 1400 AD. However, the data suggest that field strength is still high ( 65 mT) around 1150 AD (confirmed by the new result of site 50025A) and that an intensity peak occurred around 1300 AD (described from three well-dated archaeointensities from central Spain (Go´mez-Paccard et al., 2008)). The western European archaeointensities (Fig. 7a) indicate that abrupt geomagnetic field intensity changes took place in this region during the last millennia. If we consider mean intensities and ages described in Table 2, intensity changes as fast as 18 and 12 mT/century are obtained between 622.5 AD (site 41176-9) and 750 AD (site 36137A) and between 787.5 AD (site TM-04þTM13) and 945 AD (site 50353), respectively. Higher rates of change, of about 50 mT/century, are obtained from mean values of sites 36137A (750 AD) and TM-03þ06 þ07 þ09 (775 AD). Differences between the mean values available for the 12th century AD indicate a rapid intensity change of 25 mT/century (Fig. 7a).

141

Finally, an extreme variation of geomagnetic field strength of 80 mT/century during the second half of the 13th century AD is deduced from sites A08 (Genevey and Gallet, 2002) and CALA (Go´mez-Paccard et al., 2008). This rate is calculated from data obtained from Thellier derived experiments including pTRM checks and TRM anisotropy and cooling rate corrections from the study of very well dated potsherds (A08, 1275 725 AD) and one archaeological kiln abandoned between 1275 and 1300 AD (CALA). We must notice, however, that age uncertainties and errors on intensity estimations can affect considerably the calculated rates of change. For example, if we consider the minimum possible age for site A08 and the maximum for CALA, a lower rate of 20 mT/century is obtained. Nevertheless, the results indicate that the rates of intensity changes that took place in western Europe during the last millennia are much higher than has been previously described (about 7 mT per century) (Genevey and Gallet, 2002; Genevey et al., 2009). However, they are lower than the rates of change of more than 100 mT/century described by Shaar et al. (2011) from Iron Age copper slag deposits from Israel. The Bayesian approach (Lanos 2004; Lanos et al., 2005) has been applied to calculate the geomagnetic field intensity PSV curve using the aforementioned dataset (Fig. 7c). The low number of data available between 500 and 1000 AD and the offset between the two intensities available around 600 AD are clearly reflected in the results obtained. The Bayesian curve presented here must be, therefore, considered as a preliminary result. Future improvements in the number of well-dated source data (especially for the period 500–800 AD and for ages around 1150 and 1300 AD) will allow a robust description of geomagnetic field intensity variations and of the rates of intensity change and, therefore, will improve our knowledge of geomagnetic field behaviour. 6.3. Geomagnetic field intensity changes in eastern Europe between 200 and 1400 AD In order to investigate if the intensity maxima observed in western Europe are seen at a continental scale we have selected high-quality archaeointensities from Bulgaria, Italy and Greece with ages ranging from 200 to 1400 AD. As for western Europe only data derived from Thellier derived experiments and including pTRM checks have been retained. A total of 42 data from Bulgaria (N ¼33, Kovacheva et al., 2009a), Greece (N ¼6, De Marco et al., 2008; Spatharas et al., 2000, 2011; Tema et al., 2012) and Italy (N ¼3; Leonhardt et al., 2006; Tema et al., 2010) have been selected and relocated to the latitude of Sofia through the VADM (Fig. 7b). The selected data and the derived Bayesian curve (Fig. 7d) indicate very similar intensity changes to those observed in western Europe, with the occurrence of two intensity bumps (up to 75 mT) at ages around 650 and 950 AD. The results also suggest the occurrence of two periods of rapid intensity changes between 1100 and 1200 AD and around 1300 AD. The results indicate that extreme intensity variation as fast as 20 mT per century occurred in eastern Europe during historical times (Fig. 7b). The intensity pattern described (Fig. 7b and d) is similar to the one obtained from Bulgarian data using a different technique for computation of the mean intensity curve (see Fig. 10, Genevey et al., 2003). It must be underlined that the two independent local datasets treated (western and eastern Europe) show very similar features with (probably) two maxima around 600–650 AD, 800–950 AD, and two abrupt geomagnetic field intensity changes during the 12th century AD and around the second half of the 13th century AD (Fig. 7a–d). This indicates that these sharp geomagnetic field intensity changes occurred at the continental scale. The sharp increase that took place in Europe around 800–950 AD is well

142


documented in both areas whereas the occurrence of a sharp intensity peak around 600–650 AD must be confirmed for western Europe (it is only described by one data). Finally, the rapid intensity variation that probably took place in eastern Europe during the second half of the 13th century AD needs also to be confirmed. The occurrence of a double oscillation of geomagnetic field intensity in Europe has previously been suggested by Pesonen and Leino (1998) from the study of Finish archaeointensity data (with the two intensity maxima centred at 500 and 800 AD), suggesting that the double oscillation described here is also observed at more northerly latitudes. Finally, the western and the eastern selected data were compared with SCHA.DIF.3k regional (Pavoń-Carrasco et al., 2009) and ARCH3 K.1 global (Korte et al., 2009) geomagnetic field models (Fig. 7e and f). The SCHA.DIF.3K model results are in close agreement with the western European data (except for the new data obtained around 600 AD) but do not reproduce well the eastern European dataset. This regional model also suggests the occurrence of relative minor intensity maxima around 600 AD. The ARCH3K.1 global model seems to reproduce similar variations to those predicted by the regional model but gives lower intensities for the period of high intensities observed in western Europe around 800 AD (Fig. 7e). Compared with the eastern European data both models do not fit well the two periods of high intensities observed (Fig. 7f). It is worth noting that these models are based on all the available archaeointensity data without any pre-selection based on their reliability. Therefore, model results may be affected by some inaccurate data presented in the databases. In our opinion, these results indicate that a selection of archaeointensity data, based on the technique used for archeointensity determination, should be performed if a higher resolution description of geomagnetic field intensity changes is to be obtained.

7. Conclusion Ten high-quality archaeointensities from the study of 13 precisely dated early medieval Spanish pottery fragments, four French kilns and three collections of bricks used for the construction of French historical buildings with ages ranging between 335 and 1260 AD have been obtained. The new intensities – corrected for TRM anisotropy and cooling rate effects – can be considered as reliable markers of past geomagnetic field intensity. Together with a selection of the previous data that we deem to be the most reliable for western Europe, they enable an improved description of geomagnetic field intensity changes that took place in this region between the 3rd and the 14th centuries AD. The results confirm that an intensity maximum of 85 mT (at the latitude of Paris) occurred in western Europe at around 800 AD and suggest that a previous abrupt intensity change took place around 600 AD. Western European data also suggest the occurrence of abrupt geomagnetic field intensity changes during the 12th century AD and around the second half of the 13th century AD. Reliable selected eastern European data show a similar variation of geomagnetic field intensity with the occurrence of two intensity bumps (up to 75 mT at the latitude of Sofia) at ages around 650 and 950 AD and two periods of rapid intensity changes during the 12th century AD and 1300 AD. The results suggest that the described features of the geomagnetic field are observed at a continental scale and that very rapid intensity changes (of at least of 20 mT/century) took place in the recent history of the Earth’s magnetic field. The results call for additional precisely dated archaeointensities and for a selection of high-quality data, for

obtaining a refined path of the evolution of the Earths magnetic field intensity at regional scales.

Acknowledgements Financial support was given by a CSIC JAE-Doc contract (MGP), by the Nos. CGL2008-02203/BTE, CGL2011-24790, and CGL201015767/BTE projects from the Spanish ministry of Science and Innovation, the Complutense University, the Comunidad de Madrid Consolidacioń de grupos program (No. 910396), and by the Junta de Comunidades de Castilla-La-Mancha. M.K. acknowledges Geósciences-Rennes (Universite´ de Rennes 1), where she has spent two months working at the palaeomagnetic laboratory. We thank the editor and one anonymous referee for their valuable comments that greatly improved this paper. The authors acknowledge their European ancestors for their pottery, bricks and kilns manufacture.

Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.epsl.2012.08.037.

References Bard, E., Delaygue, G., 2008. Comment on ‘‘Are there connections between the Earth’s magnetic field and climate?’’ by V. Courtillot et al. [Earth and Planetary Science Letters, 253, 328, 2007]. Earth Planet. Sci. Lett. 265, 302–307, http://dx .doi.org/10.1016/j.epsl.2007.09.046. Biggin, A.J., 2010. Are systematic differences between thermal and microwave Thellier-type palaeointensity estimates a consequence of multidomain bias in the thermal results? Phys. Earth Planet. Inter. 180, 16–40, http://dx.doi.org/ 10.1016/j.pepi.2010.03.005. Catanzariti, G., Go´mez-Paccard, M., McIntosh, G., Pavoń-Carrasco, F.J., Chauvin, A., Osete, M.L., 2012. New archaeomagnetic data recovered from the study of Roman and Visigothic remains from central Spain (3rd–7th centuries). Geophys. J. Int. 188, 979–993, http://dx.doi.org/10.1111/j.1365-246X.2011.05315.x. Chauvin, A., Garcia, Y., Lanos, P., Laubenheimer, F., 2000. Paleointensity of the geomagnetic field recovered on archaeomagnetic sites from France. Phys. Earth Planet. Inter. 120, 111–136. Coe, R.S., 1967. The determination of paleointensities of Earth’s magnetic field with emphasis on mechanisms which could cause non ideal behavior in Thellier’s method. J. Geomagn Geoelectr. 19, 157–180. Coe, R.S., Gromme´, C.S., Mankinen, E.A., 1978. Geomagnetic paleointensities from radiocarbon-dated lava flows on Hawaii and the question of the Pacific nondipole low. J. Geophys. Res. 83, 1740–1756. Courtillot, V., Gallet, Y., Le Mouel, J.L., Fluteau, F., Genevey, A., 2007. Are there connections between the Earth’s magnetic field and climate? Earth Planet. Sci. Lett. 253, 328–339, http://dx.doi.org/10.1016/j.epsl.2006.10.032. De Marco, E., Spatharas, V., Go´mez-Paccard, M., Chauvin, A., Kondopoulou, D., 2008. New archaeointensity results from archaeological sites and variation of the geomagnetic field intensity for the last 7 millennia in Greece. Phys. Chem. Earth 33, 578–595, http://dx.doi.org/10.1016/j.pce.2008.02.025. Donadini, F., Kovacheva, M., Kostadinova, M., Hedley, I.G., Pesonen, L.J., 2008. Palaeointensity determination on an early medieval kiln from Switzerland and the effect of cooling rate. Phys. Chem. Earth 33, 449–457, http://dx.doi.org/ 10.1016/j.pce.2008.02.019. Finlay, C.C., 2008. Historical variation of the geomagnetic axial dipole. Phys. Earth Planet. Inter. 170, 1–14, http://dx.doi.org/10.1016/j.pepi.2008.06.029. Fisher, R.A., 1953. Dispersion on a sphere. Proc. R. Soc. London A 217, 295. Gallet, Y., Genevey, A., Le Goff, M., 2002. Three millennia of directional variation of the Earth’s magnetic field in Western Europe as revealed by archaeological artefacts. Phys. Earth Planet. Inter. 131, 81–89. Gallet, Y., Genevey, A., Courtillot, V., 2003. On the possible occurrence of archaeomagnetic jerks in the geomagnetic field over the past three millennia. Earth Planet. Sci. Lett. 214, 237–242. Gallet, Y., Genevey, A., Fluteau, F., 2005. Does Earth’s magnetic field secular variation control centennial climate change? Earth Planet. Sci. Lett. 236, 339–347. Gallet, Y., Genevey, A., Le Goff, M., Warme´, N., Gran-Aymerich, J., Lefevre, A., 2009a. On the use of archeology in geomagnetism, and vice-versa: recent developments in archeomagnetism. C. R. Phys. 10, 630–648. Gallet, Y., Hulot, G., Chulliat, A., Genevey, A., 2009b. Geomagnetic field hemispheric asymmmetry and archeomagnetic jerks. Earth Planet. Sci. Lett. 284 (1– 2), 179–186, http://dx.doi.org/10.1016/J.EPSL.2009.04.28.


Genevey, A., Gallet, Y., 2002. Intensity of the geomagnetic field in western Europe over the past 2000 years: new data from ancient French pottery. J. Geophys. Res. 107, 2285, http://dx.doi.org/10.1029/2001JB000701. Genevey, A., Gallet, Y., Margueron, 2003. Eight thousand years of geomagnetic field intensity variations in the eastern Mediterranean. J. Geophys. Res. 108, 228, ht tp://dx.doi.org/10.1029/2001JB001612. Genevey, A., Gallet, Y., Constable, C.G., Korte, M., Hulot, G., 2008. ArcheoInt: an upgraded compilation of geomagnetic field intensity data for the past ten millennia and its application to the recovery of the past dipole moment. Geochem. Geophys. Geosyst. 9 (4), Q04038, http://dx.doi.org/10.1029/ 2007GC001881. Genevey, A., Gallet, Y., Rosen, J., Le Goff, M., 2009. Evidence for rapid geomagnetic field intensity variations in Western Europe over the past 800 years from new French archeointensity data. Earth Planet. Sci. Lett. 284, 132–143, http://dx.do i.org/10.1016/j.epsl.2009.04.024. Gilks, W.R., Richardson, S., Spiegelhalter, D.J., 1996. Markov Chain Monte Carlo in Practice, Interdisciplinary Statistics. Chapman and Hall, London. ˜ ez, J.I., Go´mez-Paccard, M., Catanzariti, G., Ruiz-Martinez, V.C., McIntosh, G., Nuń Osete, M.L., Chauvin, A., Lanos, P., Tarling, D.H., Bernal-Casasola, D., Thiriot, J., Archaeological Working group, 2006a. A catalogue of Spanish archaeomagnetic data. Geophys. J. Int. 166, 1125–1143, http://dx.doi.org/10.1111/j.1365246X.2006.03020.x. Go´mez-Paccard, M., Chauvin, A., Lanos, P., McIntosh, G., Osete, M.L., Catanzariti, G., ˜ ez, J.I., 2006b. First archaeomagnetic secular variation Ruiz-Martinez, V.C., Nuń curve for the Iberian Peninsula: comparison with other data from western Europe and with global geomagnetic field models. Geochem. Geophys. Geosyst. 7, Q12001, http://dx.doi.org/10.1029/2006GC001476. Go´mez-Paccard, M., Chauvin, A., Lanos, P., Thiriot, J., Jimenez-Castillo, P., 2006c. Archeomagnetic study of seven contemporaneous kilns from Murcia (Spain). Phys. Earth Planet. Inter. 157, 16–32, http://dx.doi.org/10.1016/j.pepi.2006. 03.001. Go´mez-Paccard, M., Chauvin, A., Lanos, P., Thiriot, J., 2008. New archeointensity data from Spain and the geomagnetic dipole moment in western Europe over the past 2000 years. J. Geophys. Res. 113, B09103, http://dx.doi.org/10.1029/ 2008JB005582. Go´mez-Paccard, M., McIntosh, G., Chauvin, A., Beamud, E., Carrasco, F.J., Thiriot, J., 2012. Archaeomagnetic and rock magnetic study of six kilns from North Africa (Tunisia and Morocco). Geophys. J. Int. 189, 169–186, http://dx.doi.org/ 10.1111/j.1365-246X.2011.05335.x. Gubbins, D., Jones, A.L., Finlay, C.C., 2006. Fall in Earth’s magnetic field is erratic. Science 312, 900–902, http://dx.doi.org/10.1126/science.1124855. Hartmann, G.A., Genevey, A., Gallet, Y., Trindade, R.I.F., Le Goff, M., Najjar, R., Etchevarne, C., Afonso, M.C., 2011. New historical archeointensity data from Brazil: evidence for a large regional non-dipole field contribution over the past few centuries. Earth Planet. Sci. Lett. 306, 66–76, http://dx.doi.org/10.1016/ j.epsl.2011.03.030. Herve´, G., Schnepp, E., Chauvin, A., Lanos, P., Nowaczyck, N., 2011. Archaeomagnetic results on three Early Iron Age salt-kilns from Moyenvic (France). Geophys. J. Int. 185, 144–156, http://dx.doi.org/10.1111/j.1365-246X.2011.01933.x. Jackson, A., Jonkers, A.R.T., Walker, M.R., 2000. Four centuries of geomagnetic secular variation from historical records. Philos. Trans. R. Soc. London, A 358 (1768), 957–990, http://dx.doi.org/10.1098/rsta.2000.0569. Kirschvink, J.L., 1980. The least-squares line and plane and the analysis of paleomagnetic data. Geophys. J. R. Astron. Soc. 62, 699–718. Korte, M., Donadini, F., Constable, C.G., 2009. Geomagnetic field for 0–3 ka: 2. A new series of time-varying global models. Geochem. Geophys. Geosyst. 10, Q06008, http://dx.doi.org/10.1029/2008GC002297. Kovacheva, M., Boyadziev, Y., Kostadinova-Avramova, M., Jordanova, N., Donadini, F., 2009a. Updated archeomagnetic data set of the past 8 millennia from the Sofia laboratory, Bulgaria. Geochem. Geophys. Geosyst. 10, Q05002, http://dx. doi.org/10.1029/2008GC002347. Kovacheva, M., Chauvin, A., Jordanova, N., Lanos, P., Karloukovski, V., 2009b. Remanence anisotropy effect on the palaeointensity results obtained from various archaeological materials, excluding pottery. Earth Planets Space 61, 711–732. Lanos, P., 2001. L’approche baye´sienne en chronome´trie: application a l’archeómagne´tisme. In: Barrandon, J.N., Guibert, P., Michel, V. (Eds.), Datation, XXIe rencontres internationales d’archeólogie et d’histoire d’Antibes. e´ditions APDCA, Antibes, pp. 113–139. Lanos, P., 2004. Bayesian inference of calibration curves: application to archaeomagnetism. In: Buck, C.E., Millard, A.R. (Eds.), Tools for Constructing Chronologies: Crossing Disciplinary Boundaries. Lecture Notes in Statistics. Springer, London, pp. 43–82. Lanos, P., Le Goff, M., Kovacheva, M., Schnepp, E., 2005. Hierarchical modeling of archaeomagnetic data and curve estimation by moving average technique. Geophys. J. Int. 160, 440–476.

143

Leonhardt, R., Matzka, J., Nichols, A.R.L., Dingwell, D.B., 2006. Cooling rate correction for paleointensity determination for volcanic glasses by relaxation geospeedometry. Earth Planet. Sci. Lett. 243, 282–292, http://dx.doi.org/ 10.1016/j.epsl.2005.12.038. Ma´rton, P., 2010. Two thousand years of geomagnetic field direction over central Europe revealed by indirect measurements. Geophys. J. Int. 181, 261–268. Muscheler, R., Joos, F., Beer, J., Muller, S.A., VonMoos, M., Snowball, I., 2007. Solar activity during the last 1000 yr inferred from radionuclide records. Q. Sci. Rev. 26, 82–97. Pavoń-Carrasco, F.J., Osete, M.L., Torta, J.M., Gaya-Pique, L.R., 2009. A regional archeomagnetic model for Europe for the last 3000 years, SCHA.DIF.3K: applications to archeomagnetic dating. Geochem. Geophys. Geosyst. 10 (22), Q03013, http://dx.doi.org/10.1029/2008GC002244. Pesonen, L.J., Leino, A.H., 1998. Archaeomagnetic intensity in Finland and Bulgaria—problems in the applied technique or evidence of a non-dipole field? Phys. Chem. Earth 23, 1021–1025. Pick, T., Tauxe, L., 1993. Holocene paleointensities: Thellier experiments on submarine basaltic glass from the East Pacific Rise. J. Geophys. Res. 98, 17949–17964. Pre´vot, M., Mankinen, E.A., Coe, R.S., Gromme´, C.S., 1985. The Steens Mountain (Oregon) geomagnetic polarity transition. 2. Field intensity variations and discussion of reversal models. J. Geophys. Res. 90, 10417–10448. RenDateModel, Lanos, P., Dufresne, P., /http://www.projet-plume.org/fr/relier/ rendatemodelS. ¨ Sapin, C., Bayle´, M., Buttner, S., Gibert, P., Blain, S., Lanos, P., Chauvin, A., Dufresne, P., Oberlin, C., 2008. Archeólogie du bâti et archeóme´trie au Mont-SaintMichel, nouvelles approches de Notre-Dame-Sous-Terre. Archeólogie Me´die´vale 38, 71–122. Schnepp, E., Lanos, P., 2005. Archaeomagnetic secular variation in Germany during the past 2500 years. Geophys. J. Int. 163, 479–490. Schnepp, E., Lanos, P., Chauvin, A., 2009. Geomagnetic paleointensity between ¨ 1300 and 1750 A.D. derived from a bread oven floor sequence in Lubeck, Germany. Geochem. Geophys. Geosyst. 10, Q08003, http://dx.doi.org/10.1029/ 2009GC002470. Shaar, R., Ben-Yosef, E., Ron, H., Tauxe, L., Agnon, A., Kessel, R., 2011. Geomagnetic field intensity: how high can it get? How fast can it change? Constraints from Iron Age copper slag. Earth Planet. Sci. Lett. 301, 297–306. Shcherbakova, V.V., Shcherbakov, V.P., Heider, F., 2000. Properties of partial thermoremanent magnetization in PSD and MD magnetite grains. J. Geophys. Res. 105, 767–781. Snowball, I., Muscheler, R., 2007. Palaeomagnetic intensity data: an Achilles heel of solar activity reconstructions. Holocene 17, 851–859. Spassov, S., Hus, J., Geeraerst, R., Heller, F., 2008. Archaeomagnetic dating of a High Middle Age likely iron working site in Corroy-le-Grand (Belgium). Phys. Chem. Earth 33, 544–556. Spatharas, V., Kondopoulou, D., Liritzis, I., Tsokas, G.N., 2000. Archaeointensity results from two ceramic kilns from N. Greece. J. Balkan Geophys. Soc 3, 67–72. Spatharas, V., Kondopoulou, D., Aidona, E., Efthimiadis, K.G., 2011. New magnetic mineralogy and archaeointensity results from Greek kilns and baked clays. Stud. Geophys. Geod. 55, 131–157. Stark, F., Cassidy, J., Hill, M., Shaw, J., Sheppard, P., 2010. Establishing a first archeointensity record for the SW Pacific. Earth. Planet. Sci. Lett. 298, 113–124. Suttie, N., Holme, R., Hill, M.J., Shaw, J., 2011. Consistent treatment of errors in archaeointensity implies rapid decay of the dipole prior to 1840. Earth Planet. Sci. Lett. 304, 13–21. Tema, E., Hedley, I., Lanos, P., 2006. Archaeomagnetism in Italy: a compilation of data including new results and a preliminary Italian secular variation curve. Geophys. J. Int. 167, 1160–1171. Tema, E., Goguitchaichvili, A., Camps, P., 2010. Archaeointensity determinations from Italy: new data and the Earth’s magnetic field strength variation over the past three millennia. Geophys. J. Int. 180, 596–608, http://dx.doi.org/10.1111/ j.1365-246X.2009.04455.x. Tema, E., Go´mez-Paccard, M., Kondopoulou, D., Almar, Y., 2012. Intensity of the Earth’s magnetic field in Greece during the last five millennia: new data from Greek pottery. Phys. Earth Planet. Inter. 202–203, 14–26, http://dx.doi.org/ 10.1016/j.pepi.2012.01.012. Thellier, E., Thellier, O., 1959. Sur l’intensite´ du champ magne´tique terrestre dans le passe´ historique et geólogique. Ann. Geóphys. 15, 285–376. Wollin, G., Ericson, D.B., Ryan, W.B.F., 1978. Climatic changes, magnetic intensity variations and fluctuations of eccentricity of Earth’s orbit during past 2,000,000 years and mechanism which may be responsible for relationship. Earth Planet. Sci. Lett. 41, 395–397. Zananiri, I., Batt, C.M., Lanos, P., Tarling, D.H., Linford, P., 2007. Archaeomagnetic secular variation in the UK during the past 4000 years and its application to archaeomagnetic dating. Phys. Earth Planet. Inter. 160, 97–107.

1. Archaeological setting and dating of the early medieval Spanish pottery fragments 13 pottery fragments collected from El Tolmo de Minateda archaeological site (TM in Fig. 1) have been studied. The archaeological excavations together with toponymic studies, material and historic sources have given spectacular remains of an important city founded by the Visigothic Kingdom towards the end of the 6th century, beginning of the 7th century AD. These urban remains were found above those of a Roman settlement mainly abandoned during the Early Imperial Roman period. The city was a new episcopate created in order to administrate the ilicitan territories recently conquered by the Byzantines (Gutiérrez-Lloret, 2000; Gutiérrez-Lloret et al., 2004).

Archaeological, historical (Abad-Casal et al., 2007; Alba-Calzado and Gutiérrez-Lloret, 2008; Cañavate et al., 2009) and numismatic considerations (Doménech and GutiérrezLloret, 2006; Gutiérrez-Lloret and Doménech, in press) place the activity of the monumental visigothic complex during the 7th century and probably the beginning of the 8th century AD. The complex was abandoned towards the 8th century and was partially destroyed at around 750 AD. The subsequent Islamic occupation established a new urban organization with new utilities for the different spaces and buildings.

Excavations of the stratigraphic sequence yielded three horizons (Amorós et al., in press; Gutiérrez-Lloret et al., 2003). Horizon I, visigothic, dates from the second half of the 7th century to the first quarter of the 8th century. Unfortunately, no ceramic samples have been collected from this horizon. Horizon II, to which the majority of the studied fragments belong, corresponds to transformation of the monumental complex related to the Islamic occupation of the site and dates from the second half of the 8th century to the beginning of the 9th century. Horizon III, dates from the beginning to the end of the

1

9th century. A total of 13 well-dated ceramic fragments, 10 from Horizon II and 3 from horizon III, have been collected (Table 1, Supplementary Material Table S1). The archaeological/historical context (Amorós et al., in press; Gutiérrez-Lloret et al., 2003) together with numismatic evidence (Doménech and Gutierrez-Lloret, 2006; GutiérrezLloret and Doménech, in press) and stratigraphic constraints between the different levels found within the horizons (Amorós and Cañavate, 2010; Cañavate et al., 2009; Gutiérrez-Lloret, 2002, 2011; Gutierrez-Lloret and Cánovas, 2009) allow the precise dating of the studied ceramic fragments (see Table S1 for further details).

2. Archaeological setting and dating of bricks used for the construction of historical French buildings 2.1. Mont-Saint-Michel, Notre-Dame-Sous-Terre church (site 50353) Eight bricks were sampled from different masonries at Notre-Dame-Sous-Terre church owning to building phase 1 (bricks sampled in windows of the church) and gave very consistent thermoluminescence (TL) dates (Blain et al., 2007, Sapin et al., 2008; Table S2). Moreover, three radiocarbon dating experiments have been carried out onto charcoals recovered in mortars between bricks. Calibrated radiocarbon results are very close to TL results (Table S2). The synthetic dating interval for the firing of the bricks, obtained by Bayesian treatment applied to the eight TL and three

14

C dates, is [900,

990] AD at 95% confidence level.

2.2. Angers, Collegial church St. Martin (site 49007I) TL analyses were performed on five bricks from arches of the bell-tower of the church. Additionally, one calibrated 14C age was obtained from the study of charcoals from the

2

masonry related to bricks used for TL (Table S2). Altogether the results have been treated by Bayesian modelling and indicate that the age of this part of the church is inside the interval [810, 910] AD at 95% confidence level (Blain et al., 2011).

2.3. Castle of Avranches (site 50025A) Five bricks for TL dating were drilled from the northeast tower of the keep (Bouvier et al., 2011; Table S2). The Bayesian treatment applied to TL results indicate that the construction of the castle keep was within the interval [1100-1260] AD at 95% confidence level.

3. Archaeological setting and dating of the French archaeological kilns 3.1. Clermont-l´Hérault (site 34079C) The imported and local ceramic artefacts found in this site correspond to the end of the Roman period (Pomarèdes, 2005). This suggests that the Terminus Post Quem and Terminus Ante Quem of the kiln (number 4002) can be established at 300 AD and 600 AD, respectively. This archaeological prior dating interval can be modeled as an uniform interval [300, 600] AD. A radiocarbon date (LY 10772, radiocarbon Lab. of Lyon, France) has also been obtained from charcoals discovered in a kiln (number 4008) lying just near the one studied here by archaeomagnetism. Archaeological evidences indicate that these kilns are contemporaneous at some years close. The

14

C

date obtained is 1635 ± 30 yr BP, hence the calibrated interval [340, 530] AD at 95% confidence level. The final dating for the studied kiln has been deduced from these two independent dating results assuming that they correspond to the same event. We obtain

3

the synthetic interval [335, 540] AD at 95% confidence level as the final result (Table 1).

3.2. Pierrefitte-sur-Sauldre (site 41176-9) The local ceramics produced in this kiln are dated to the beginning of the 7th century AD, hence the prior dating uniform interval [600, 650] AD (Bouillon, in press; Riquier, 1995). In order to get more chronological constraints, a radiocarbon date (LY 8047, Radiocarbon Lab. of Lyon, France) has been obtained from charcoals discovered in the kiln itself. The interval 1455 ± 55 yr BP ([440, 670] calibrated yr AD) at 95% confidence level has been obtained. Altogether, the two independent dating results indicate that the last heating of the kiln occurred inside the interval [590, 655] AD at 95% confidence level.

3.3. Néons-sur-Creuse (site 36137A) The kiln was empty and cleaned, without any artefacts inside nor charcoals. However, local ceramics (cooking wares) dating from the end of Early Middle Age (7th-9th centuries AD) have been discovered inside several pits (silos) just near the kiln (Hamon, 1992). Therefore, the uniform prior dating interval [600, 900] AD can be considered for the last heating of the kiln.

3.4. Saran, ZAC des Vergers (site 45302A) The pottery workshop of Saran was occupied during a long time period, from the 6th to the end of the 9th centuries AD (Jesset, 2001, in press). In the literature we can found previous archaeomagnetic studies performed in Saran (Thellier, 1981). They correspond to “Lac de la Médecinerie” site (kilns B=110 and C=113 sampled in 1969, and kiln

4

G=131 sampled in 1971). They were attributed to the period 750 to 900 AD and were integrated in the reference French dataset (Bucur, 1994; Gallet et al., 2002) used to construct the directional PSV French curve.

In 1999-2000, new excavations directed by S. Jesset have been performed in the commune of Saran, site of ZAC des Vergers, located 2 km far from the “Lac de la Médecinerie” site studied by Thellier. Four potter’s kiln were discovered, including the kiln studied here (F256). The ceramics recovered inside this kiln are dated between the end of the 7th and the beginning of the 8th centuries AD, by comparison with ceramics discovered in well-dated sites excavated in Orleans (dated by stratigraphy, coins, small artefacts and historical texts; Jesset, 2001, in press). Therefore, from archaeological evidences, the dating interval for the kiln studied here can be established at [675, 725] AD.

References Abad-Casal, L., Cánovas, P., Gamo, B., Gutiérrez-Lloret, S., 2007. El complejo episcopal de Eio (el Tolmo de Minateda, Albacete). Últimas aportaciones arqueológicas (Cuenca 2005). Arqueología de Castilla-La Mancha. Actas de las I Jornadas, 171-185. Alba-Calzado, M., Gutiérrez-Lloret, S., 2008. Las producciones de transición al Mundo Islámico: el problema de la cerámica paleoandalusí (siglos VIII y IX), in: Bernal, D., Ribera, A. (Eds.), Cerámicas hispanorromanas: un estado de la cuestión, Cádiz, 585-613. Amorós, V., Cañavate, V., 2010. Transformación funcional de espacios representativos en los inicios del emirato. La basílica y el espacio episcopal de El Tolmo de

5

Minateda, I Congreso Internacional, Espacios urbanos en el Occidente Mediterráneo (ss. VI-VIII), Toledo, 191-198. Amorós, V., Cañavate, V., Gutiérrez-Lloret, S., Sarabia, J., in press. Cerámica altomedieval en el Tolmo de Minateda (Hellín, Albacete, España), IX Congreso Internazionale AIECM2, Venezia. Blain S., Guibert, P., Bouvier, A., Vieillevigne, E., Brechtel, F., Sapin, C., Baylé, M., 2007. TL-dating applied to building archaeology: The case of the medieval church

Notre-Dame-Sous-Terre

(Mont-Saint-Michel,

France).

Radiation

Measurements 42, 1483-1491. Blain S., Guibert, P., Prigent, D., Lanos, P., Oberlin, C., Sapin, C., Bouvier, A., Dufresne, P., 2011. Combined dating methods applied to building archaeology: the contribution of Thermoluminescence to the case of the bell tower of the St. Martin’s

Church,

Angers

(France).

Geochronometria

38,

55-63.

doi:10.2478/s13386-011-0010-0. Bouillon, J., in press. Les cadres de productions potiers du haut Moyen Âge en region centre: Etat des lieux, Actes du colloque de Douais, Tourner autour du pot : les ateliers de potiers médievaux du Ve au XIIe siècle dans léspace européen. Bouvier, A., Pinto, G., Guibert, P., Nicolas-Méry, D., Baylé, M., 2011. Luminescence dating applied to medieval architecture: the north east tower of the Avranches keep (Manche, France). Archéosciences, Revue dÁrchéometrie 34, 59-68. Bucur, I., 1994. The direction of the terrestrial magnetic field in France, during the last 21 centuries. Recent progress, Phys. Earth Planet. Inter. 87, 95-109. Cañavate, V., Mellado, J.A., Sarabia, J., 2009. Uso, residualidad y la problemática del siglo VIII en el palacio visigodo del Tolmo de Minateda (Hellín, Albacete). Arqueología y Territorio Medieval 16, 9-32.

6

Doménech, C., Gutiérrez-Lloret, S., 2006. Viejas y nuevas monedas en la ciudad emiral de Madinat Iyyuh (El Tolmo de Minateda, Hellín, Albacete). Al-Qantara XXVII 2, 337-374. Gallet, Y., Genevey, A., Le Goff, M., 2002. Three millennia of directional variation of the Earth’s magnetic field in Western Europe as revealed by archaeological artefacts. Phys. Earth Planet. Inter. 131, 81-89. Gutiérrez-Lloret, S., 2000. La identificación de Madinat Iyih y su relación con la sede episcopal Elotana: Nuevas perspectivas sobre viejos problemas. Script ain Honorem E. A. Llobregat, Alicante, 481-501. Gutiérrez-Lloret, S., 2002. De espacio religioso a espacio profano: transformación del área urbana de la basílica del Tolmo de Minateda (Hellín, Albacete) en barrio islámico, II Congreso de Historia de Albacete (Albacete, Noviembre del 2000). Gutiérrez-Lloret, S., 2011. El Tolmo de Minateda en torno al 711. Zona Arqueológica 15, 353-372. Gutiérrez-Lloret, S., Cánovas, P., 2009. Construyendo el siglo VII: arquitectura y sistemas constructivos en el Tolmo de Minateda, El siglo VII frente al siglo VII. Arquitectura, Anejos de AEspA XLVIII, 91-131. Gutiérrez-Lloret, S., Doménech, C., in press. Monedas en contexto: la ciudad altomedieval de El Tolmo de Minateda (Hellín, Albacete, España), Preatti del I Workshop Internazionale di Numismatica, Roma, 28-30. Gutiérrez-Lloret, S., Gamo, B., Amorós, V., 2003. Los contextos cerámicos altomedievales del Tolmo de Minateda y la cerámica altomedieval en el sureste de la Península Ibérica. In: Caballero, L. Mateos, P., Retuerce, M. (Eds.), Cerámicas tardorromanas y altomedievales en la Península Ibérica: ruptura y

7

continuidad (II Simposia de Arqueología, Mérida 2001), Anejos de AEspA XXVIII, Madrid, pp. 119-168. Gutiérrez-Lloret, S., Abad-Casal, L., Gamo, B., 2004. La iglesia visigoda de El Tolmo de Minateda (Hellín, Albacete), Sacralidad y Arqueología. In: Blázquez, J.M., Gónzalez-Blanco, A. (Eds.), Thilo Ulbert zum 65 Geburtstag am Juni 2004 gewidmet, Antigüedad y Cristianismo (Murcia), XXI, 137-170. Hamon, T., 1992. Néons-sur-Creuse, les Cognées. In : Bilan scientifique de la Région Centre, Service Régional de l’Archéologie, Direction du Patrimoine, Orléans, 67-69. Jesset, S., 2001. Saran (Loiret), ZAC des Vergers, site 45 302 08 AH : Rapport de fouille. Document final de synthèse (DFS). Inédit. Service Régional de l’Archéologie de la Région Centre, Orléans, 267 p + annexes. Jesset, S., in press. Les ateliers de potiers du Haut Moyen Âge autour dÓrléans : caractérisation, organisation et production, Actes du colloque de Douais, Tourner autour du pot : les ateliers de potiers médievaux du Ve au XIIe siècle dans léspace européen. Pomarèdes, H., 2005. La Quintarié (Clermont-l’Hérault, 34): établissement agricole et viticulture, atelier de céramiques paléochrétiennes (DS.P), Ier-VI s. ap. J.C., éditions Monique Mergoil, Montagnac, 193. Riquier, S., 1995. Découverte dún atelier de potier mérovingien à Pierrefitte-surSauldre (Loir-et-Cher). In : Deletang, H. (Ed.). Archéologie en Sologne. Bulletin du Groupe de recherches archéologiques et historiques de Sologne 17(3-4), 159166. Sapin, C., Baylé, M., Büttner, S., Gibert, P., Blain, S., Lanos, P., Chauvin, A., Dufresne, P., Oberlin, C., 2008. Archéologie du bâti et archéométrie au Mont-Saint8

Michel,

nouvelles

approches

de

Notre-Dame-Sous-Terre.

Archéologie

Médiévale 38, 71-122. Thellier, E., 1981. Sur la direction du champ magnétique terrestre en France, durant les deux derniers millénaires. Phys. Earth Planet. Inter. 24, 89-132.

9

4. Supplementary Tables Ceramic

Archaeological and stratigraphic constraints

fragment TM-02

Age (yrs AD)

Level in the lower part of horizon III (800-900 AD). Removal of debris from the Palatial hall. A

800-850

surface in the first phase of the neighbourhood/district. This level corresponds to the first half of the 9th century. TM-03

Level from horizon II (750-800 AD). Collapse associated with the open space between the church

750-800

and the Palace. It covers the surface associated with an orange repavement dated from the mid-8th century. This level corresponds to the second half of the 8th century. TM-04

Level from horizon II (750-800 AD). Collapse associated with the Palatial complex, between the

775-800

two buildings. A surface prior to the Emiral district (beginning of the 9th century) but posterior to the orange pavement. This level correspond to the last quarter of the 8th century. TM-05

Level from horizon II (750-800 AD). Collapse of the Palatial hall, posterior to the orange

750-800

repavements, dated from the mid-8th century. TM-06

Level from horizon II (750-800 AD). Collapse of one of the Palace rooms. It covers part of the

750-800

room’s orange repavements, dated from the mid-8th century. This level corresponds to the second half of the 8th century. TM-07

Level from horizon II (750-800 AD). Part of the collapse of one of the Palace rooms. It covers

750-800

part of the room’s orange repavement, dated from the mid-8th century. This level corresponds to the second half of the 8th century. TM-08

Level from horizon II (750-800 AD). Orange pavement of one of the rooms of the Visigoth

750-800

Palace. This level corresponds to the second half of the 8th century. TM-09

Level from horizon II (750-800 AD). A heap of the first documented abandonments in one of the

750-800

Palace rooms. This level corresponds to the second half of the 8th century. TM-10

Level from horizon II (750-800 AD). Remains of the orange pavement pertaining to a room of

750-800

the Visigoth Palace. This level corresponds to the second half of the 8th century. TM-11

Level from horizon II (750-800 AD). Remains of the orange pavement pertaining to a room of

750-800

the Visigoth Palace. This level corresponds to the second half of the 8th century. TM-12

Level in the middle of horizon III (800-900 AD). Surface of an Emiral house. This level

825-875

corresponds to the middle 9th century. TM-13

Level in the upper part of horizon II (750-800 AD). A rubbish tip associated with prior Islamic

775-800

district level and the pillaging of the church. This level corresponds to the last quarter of the 8th century. TM-14

Level in the upper part of horizon III (800-900 AD). Filling/heap on the praefurnium of an Emiral

850-900

ceramic kiln. This level corresponds to the second half of the 9th century.

Table S1: Archaeological and stratigraphic constraints and dating of the stratigraphic levels where the studied ceramic fragments were found. All the fragments were recovered at El Tolmo de Minateda (Hellín, Spain) archaeological site.

10

Notre-Dame-Sous-Terre church (site 50353) Thermoluminescence results Sample (bricks)

Date (AD) +- sigma overall Mansory

Bdx8853

947 ± 131

Window US 65

Bdx8854

1049 ± 93

Window US 65

Bdx8856

947 ± 100

Window US 65

Bdx8866

941 ± 89

Window US 66

Bdx8861

946 ± 85

Window US 62

Bdx8862

962 ± 75

Window US 62

Bdx8858

916 ± 86

Window US 73

Bdx8859

832 ±168

Window US 73

Sample location

Lab Reference

Age (BP)

Calibrated date (95%)

masonry 63G

Lyon-3131 (Poz)

1100 ± 30

891-1000

masonry 63

Lyon-3130 (Poz)

1140 ± 30

782-981

masonry 24

Lyon-3127 (Poz)

1100 ± 30

888-999

14

C results

Collegial church St. Martin (site 49007I) Thermoluminescence results Sample (bricks)

Date (AD) +- sigma overall

Bdx9456

928 ± 65

Bdx9457

900 ± 79

Bdx9459

825 ± 79

Bdx9461

844 ± 82

Bdx9472

735 ± 83

14

C results

Sample location

Lab Reference

Age (BP)

Calibrated date (95%)

Arch of the bell-tower

Lyon-7857

1202 ± 34

735-942

Castle of Avranches (site 50025A) Thermoluminescence results Sample (bricks)

Date (AD) +- sigma overall

Bdx8846

1049 + 149

Bdx8847

1090 + 88

Bdx8848

1233 + 65

Bdx8849

1220 + 62

Bdx8850

1176 + 82

Table S2: Thermoluminescence and historical French buildings.

14

C dating results obtained from samples of the

11

Site

Specimen

NRM (A/m)

 -6

Tmin-Tmax

n

f

g

q

MAD

DANG

F  sd

Fe

∆TRM

alt.

(10 )

(°C)

(°)

(°)

(µT)

(µT)

(%)

(%)

Angers,

41B

1.54

477

350-550

5

0.58

0.71

8.3

2.0

0.4

89.0 4.4

89.0

8.4

-3.6

Collegial church St. Martin

42B

4.91

880

100-500

8

0.89

0.70

33.8

1.4

0.7

82.1 1.5

82.1

12.9

0.6

49007I

43B

2.47

675

100-550

9

0.94

0.85

43.6

2.2

1.4

83.6 1.6

83.5

10.8

-2.7

Bricks in buildings

44B

18.40

1371

100-550

9

0.97

0.87

65.7

2.0

0.1

82.2 1.1

79.9

9.1

0.1

45B

2.40

1034

100 550

9

0.88

0.86

27.2

3.4

1.0

78.9 2.0

78.2

13.3

-3.0

46B

6.71

1949

100-500

8

0.89

0.81

33.2

2.3

1.7

76.8 1.7

77.2

14.9

-0.7

52B

4.77

2191

100-500

8

0.93

0.85

50.8

2.2

1.1

82.3 1.3

79.5

12.2

0.9

53B

2.23

1309

100-550

9

0.90

0.86

53.3

1.9

0.8

77.1 1.1

74.4

54B

3.14

564

100-550

9

0.93

0.80

34.1

2.4

0.4

76.8 1.7

80.2

9.2

-0.4

94B

12.56

1889

100-470

8

0.88

0.81

19.0

2.8

2.0

72.6 2.7

72.8

-7.0

1.8

98B

10.20

1178

100-490

9

0.88

0.83

36.6

2.1

0.6

74.1 1.5

71.7

-9.4

2.9

102B

13.82

795

100-530

11

0.92

0.88

39.1

3.2

0.7

72.8 1.5

70.9

-10.8

-0.3

112B

11.51

413

100-560

13

0.94

0.82

3.1

2.3

0.8

74.1 19.1

74.0

-12.2

2.6

114B

3.12

540

100-530

11

0.87

0.88

32.3

2.6

1.9

86.4 2.0

93.5

-12.0

4.5

115B

3.12

837

100-530

11

0.92

0.88

5.6

5.2

2.0

77.0 11.2

74.7

-2.5

-1.8

229B

10.81

1285

100-490

9

0.86

0.84

25.6

2.6

1.3

75.2 2.1

-9.7

-0.1

230B

2.98

572

100-530

11

0.86

0.89

59.0

2.0

0.5

84.9 1.1

84.9

-21.9

13.3

236B

0.19

80

100-530

11

0.59

0.86

11.2

6.8

4.1

59.3 2.6

61.5

-18.2

5.5

242B

0.87

472

100-470

8

0.69

0.78

19.6

4.4

2.2

65.1 1.8

72.6

-13.2

2.1

244B

0.74

429

100-490

9

0.70

0.77

18.5

3.5

2.6

66.5 1.9

67.6

-16.2

4.8

248B

3.23

567

100-530

11

0.87

0.88

25.8

3.1

2.5

88.9 2.6

92.2

-5.1

-8.9

249B

3.17

659

100-530

11

0.85

0.88

94.0

3.1

1.2

91.5 0.7

93.8

-8.2

0.9

9BB

14.90

2198

100-470

8

0.81

0.83

30.9

2.6

0.4

71.8 1.6

76.1

-11.5

3.7

Avranches,

7B1

10.39

1495

275-550

12

0.88

0.84

71.4

3.5

0.1

62.4 0.7

58.1

5.2

Castle of Avranches

1B1

3.23

826

200-550

14

0.89

0.91

78.2

2.3

0.6

68.3 0.7

71.9

3.8

50025A

3B1

2.55

502

250-550

13

0.84

0.87

33.0

2.6

1.6

76.0 1.7

71.1

Bricks in buildings

5B1

3.86

1500

200-550

14

0.80

0.91

46.0

2.7

1.1

72.5 1.1

67.8

2.9

-11.0

9B1

4.26

1755

100-525

14

0.93

0.91

43.1

4.7

1.8

78.6 1.6

66.5

11.0

-6.0

2B1

4.74

736

200-500

14

0.88

0.90

54.0

1.8

0.2

77.9 1.1

74.0

6.5

-2.9

12

4B1

4.80

718

200-550

14

0.92

0.90

59.3

1.7

0.5

66.8 0.9

68.0

2.3

-5.7

6B1

2.84

775

100-550

15

0.94

0.92

77.4

1.7

1.1

57.9 0.7

68.8

4.2

-4.7

8B1

18.1

1944

300-550

11

0.92

0.84

51.1

3.1

0.8

67.3 1.0

63.7

6.4

-4.8

10B1

13.70

1975

100-550

15

0.97

0.86

52.2

2.9

1.0

71.2 1.1

70.4

10.6

-10.7

Mont-Saint-Michel,

A-2B1

21.00

1495

300-500

7

0.58

0.79

20.0

8.0

5.4

77.0 1.8

79.1

Notre-Dame-Sous-Terre church

A-3B1

8.91

1305

200-533

10

0.88

0.83

46.3

6.6

3.6

78.8 2.2

72.7

4.2

-10.1

50353

A-4B1

5.64

1138

100-533

11

0.89

0.87

26.0

4.6

1.8

86.3 2.7

78.5

5.5

-5.0

Bricks in buildings

A-5B1

7.46

880

250-533

9

0.79

0.76

29.9

6.1

1.4

60.3 1.2

61.1

2.0

-10.8

A-6B1

12.03

1389

200-533

10

0.75

0.77

27.8

6.9

0.8

50.7 1.1

55.5

4.8

-16.3

A-7B1

15.94

2680

200-566

11

0.88

0.77

87.4

4.1

1.0

80.51.0

74.8

2.3

-14.8

A-8B1

5.05

2157

200-566

10

0.66

0.85

38.5

5.3

3.1

72.2 1.5

78.7

6.1

-6.6

A-10B1

16.65

1739

200-533

11

0.79

0.75

21.8

5.9

1.3

68.5 1.4

64.0

-4.1

-10.2

B-1B1

9.37

1466

200-533

10

0.86

0.83

13.6

7.9

1.6

86.8 1.4

80.7

1.0

-11.0

B-2B3

4.05

875

100-566

12

0.95

0.90

61.6

3.0

1.6

53.3 0.8

67.4

-3.6

-10.1

B-3B1

2.66

471

100-566

12

0.94

0.87

42.5

4.0

0.2

60.5 1.9

71.6

2.2

-18.1

B-5B3

6.78

1299

200-533

10

0.88

0.87

24.7

3.0

1.1

66.2 1.4

68.7

3.0

-9.9

B-6B1

16.23

2166

200-500

9

0.53

0.83

9.0

5.6

0.9

60.2 2.9

61.6

12.8

-8.0

B-7B1

27.71

2658

200-533

10

0.85

0.81

29.9

6.3

2.1

74.7 1.8

68.7

1.6

-11.7

B-8B1

11.16

1576

200-500

9

0.60

0.82

11.3

7.7

4.1

60.3 2.6

54.6

-1.5

-10.4

D-1B

12.44

1324

200-566

11

0.94

0.83

72.8

2.6

0.2

69.8 1.2

72.9

1.0

-8.9

D-2B

4.38

806

200-566

11

0.91

0.87

50.0

3.3

1.0

65.9 1.5

70.1

5.7

-6.0

D-3B

11.43

1374

200-566

11

0.95

0.85

57.7

2.7

1.0

73.8 1.7

71.4

1.5

-7.9

D-4B

11.10

1344

200-566

11

0.91

0.81

48.3

3.7

1.7

67.9 1.5

68.7

-0.5

-10.7

D-6B

5.32

912

200-566

11

0.88

0.84

37.3

6.3

2.3

54.9 1.6

61.7

4.2

-10.1

D-7B

10.16

1197

200-533

10

0.70

0.79

18.0

4.8

1.9

68.4 2.1

62.4

6.2

3.7

D-9B

7.09

1673

100-566

12

0.91

0.89

56.6

4.8

2.3

63.2 0.9

76.1

6.4

-1.9

D-10B

3.82

619

200-566

11

0.89

0.85

41.3

2.5

1.1

82.6 2.4

70.3

10.0

11

D-11B

7.51

966

250-566

11

0.91

0.83

76.9

2.4

0.6

71.6 1.1

73.8

8.3

1.6

E-2B

7.60

1157

200-566

11

0.75

0.88

37.3

4.0

0.5

81.3 2.1

77.8

6.0

0.7

E-3B

6.86

2187

200-566

11

0.81

0.87

35.0

3.8

0.4

75.3 2.0

73.7

2.9

-4.3

E-4B

5.65

1470

200-566

11

0.82

0.87

52.4

3.3

0.8

90.5 1.4

86.2

4.4

-1.6

13

E-5B

5.81

1729

200-566

11

0.87

0.88

47.8

4.5

1.8

68.9 1.6

72.2

4.3

0.8

E-7B

3.79

1464

200-566

11

0.80

0.88

61.3

3.3

1.5

73.4 1.3

66.2

5.2

-0.8

E-8B

5.90

931

200-533

10

0.81

0.85

35.7

4.4

3.0

68.4 1.5

66.4

4.4

0.3

E-14B

2.80

341

200-533

10

0.91

0.88

40.6

2.9

1.8

71.7 1.7

77.3

9.0

2.4

E-16B

6.20

566

100-566

12

0.90

0.88

54.6

4.1

1.2

78.0 1.8

74.3

5.7

3.2

F-1B1

12.10

1617

250-566

10

0.78

0.85

28.0

8.4

1.8

71.4 3.1

74.6

2.5

-0.7

F-2B2

11.98

1457

100-566

12

0.95

0.86

35.6

6.6

1.4

71.8 1.7

73.4

4.5

0.1

F-3B2

8.83

1222

200-533

10

0.92

0.86

36.9

3.4

0.8

60.0 1.4

74.7

0

0

G-1B

4.63

674

100-533

11

0.92

0.87

37.3

4.6

3.0

77.2 1.4

74.5

3.6

5.1

Pierrefitte-sur-Sauldre

1P

7.05

2370

190-550

11

0.69

0.89

11.7

2.7

1.2

75.7 4.0

72.4

41176-9

4P

31.62

2910

190-550

11

0.71

0.85

10.8

2.5

1.6

98.6 5.6

95.9

14.1

-3.9

Kiln

1.3

6P

16.61

2230

190-550

11

0.65

0.82

8.7

2.2

1.5

98.8 4.3

98.8

11.2

Dec at Paris = 6.1º, Inc at Paris = 70.2º

8P

18.39

2940

230-550

10

0.53

0.86

7.5

2.5

0.7

93.9 5.7

94.8

11.0

1.3

n = 21, α95 = 1.1º, k = 791

9P

7.10

1270

190-550

11

0.64

0.88

13.5

2.4

1.6

100.0 4.2

97.1

20.7

32.0

19A

10.27

1480

100-510

11

0.79

0.88

11.1

2.2

2.0

97.3 6.1

96.9

0.5

1.3

8A

13.9

2481

100-510

11

0.66

0.89

22.6

2.4

1.1

110.1 2.8

91.4

6a

16.6

2065

100-510

11

0.74

0.88

17.7

1.9

1.1

98.1  1.8

96.5

Néons-sur-Creuse

8P

2.83

630

230-550

10

0.73

0.87

12.9

3.6

1.6

77.6 ± 3.8

75.0

11.3

-1.0

36137A

13P

10.13

1540

190-550

11

0.75

0.81

54.6

3.0

0.5

79.4 ± 0.9

74.8

8.2

0.0

Kiln

13A

6.09

1300

100-510

11

0.86

0.87

42.8

1.9

1.5

85.3 ± 1.5

81.4

12.3

1.4


16P

3.26

450

190-550

11

0.85

0.84

70.5

1.4

0.7

75.6 ± 0.8

74.0

17.6

3.1

n = 14, α95 = 1.6º, k = 550

17P

4.40

260

190-550

11

0.85

0.86

37.3

2.4

0.9

73.9 ± 1.4

74.0

16.2

6.4

19P

4.42

560

190-550

11

0.84

0.82

75.2

2.1

1.2

74.0 ± 0.7

74.3

13.4

1.2

19A

3.76

300

100-560

11

1.00

0.89

26.6

1.5

0.6

79.2 ± 2.7

79.1

14.3

5.8

Clermont l´Hérault

1P

3.23

500

100-620

13

0.93

0.80

115.5

3.6

1.8

73.7 ± 0.5

71.4

5.1

-1.1

34079C

12P

0.74

460

100-510

11

0.60

0.89

14.5

1.6

6.3

63.4 ± 2.3

62.1

4.1

13.6

Kiln

15P

2.00

420

100-620

13

0.98

0.83

53.2

3.4

0.8

80.4 ± 1.2

72.5

9.3

0.4

16P

3.41

1170

100-510

11

0.86

0.88

37.3

1.6

2.9

66.9 ± 1.4

59.5

-2.9

0.8

Dec at Paris = 0.4º, Inc at Paris = 62.5º n = 20, α95 = 1.1º, k = 875

14

Saran, Zac des Vergers

8P1

11.94

2060

100-560

12

0.94

0.90

57.3

1.3

3.6

79.8 ± 1.2

79.9

14

5.2

45302A

17P

5.59

700

100-560

12

0.91

0.88

67.2

1.9

1.0

84.0 ± 1.0

82.1

15.2

5.7

Kiln

9.1

3.3

18P

7.96

900

100-560

12

0.93

0.89

83.6

2.2

2.7

85.4 ± 0.8

85.3


19P

8.37

230

100-510

11

0.56

0.79

31.4

3.8

2.1

70.7 ± 1.0

69.3

n = 13, α95 = 1.9º, k = 412

20P

0.82

250

100-560

12

0.93

0.90

102.4

1.1

1.3

82.4 ± 0.7

82.0

22P

3.48

590

100-560

12

0.93

0.87

96.1

2.7

1.5

84.7 ± 0.7

84.9

23P

0.79

150

100-560

12

0.95

0.89

119.9

1.5

1.0

85.7 ± 0.6

85.7

25P

1.81

330

100-560

12

0.93

0.85

89.6

1.2

0.5

81.1 ± 0.7

80.9

26P

1.46

370

100-560

12

0.92

0.90

100.2

1.2

1.2

79.0 ± 0.7

80.2

Albacete,

02A

1.81 10-1

585

100-480

11

0.76

0.89

31.9

3.5

1.9

63.1  1.3

67.6

1.6

-0.1

Tolmo de Minateda

02B

1.34 10-1

790

150-350

5

0.46

0.74

21.4

4.2

2.9

60.9  1.0

68.9.

2.6

-0.9

TM

02C

1.88 10

-1

763

150-430

7

0.70

0.84

23.8

4.8

1.1

61.5  1.5

69.9

2.6

-1.8

Ceramic fragments

03B

6.23 10-2

171

200-470

7

0.72

0.75

41.4

3.3

1.4

64.7 0.8

64.8

-1.0

-1.1

03C

3.27 10

-2

95

300-470

5

0.61

0.71

15.5

1.8

0.5

61.9 1.7

61.4

1.5

-3.7

03D

3.78 10-2

181

250-510

7

0.87

0.79

40.3

2.9

1.8

42.3 0.7

55.8

0.8

-3.9

04A

4.41 10-1

1356

100-480

11

0.84

0.90

40.2

2.3

1.6

88.2  1.7

72.4

3.1

-1.4

04B

2.46 10

-1

1382

200-470

7

0.68

0.83

31.7

2.4

1.2

78.9  1.4

75.1

4.6

-0.5

06A

2.83 10-2

48

100-510

12

0.95

0.86

24.4

3.6

1.5

44.8  1.5

68.3

5.9

-4.2

06B

5.93 10-2

116

250-470

6

0.79

0.73

13.6

4.4

2.0

41.3  1.7

80.2

7.1

-5.2

07A

6.86 10

-1

407

330-540

8

0.84

0.73

14.4

2.5

1.3

73.4 3.2

72.7

6.4

-0.1

07C

3.18 10-1

385

300-540

7

0.86

0.70

21.2

2.1

0.8

67.9 1.9

85.0

5.8

-0.9

07D

1.99 10-1

389

350-540

6

0.83

0.68

21.8

1.6

0.6

70.5 1.8

81.1

6.0

-0.7

09A

2.51 10

-1

1737

100-450

10

0.77

0.88

31.6

4.0

1.6

61.6 1.4

64.9

6.7

0.9

09C

4.51 10-1

1432

150-430

7

0.70

0.83

24.6

2.5

1.0

64.6 1.5

71.8

1.9

-1.4

09D

2.41 10-1

1499

200-470

7

0.67

0.82

28.3

3.6

2.3

64.5 1.3

74.2

2.9

-2.9

13A

6.38 10

-1

895

100-480

11

0.80

0.88

75.3

1.7

1.3

76.4 0.7

76.9

-4.3

-1.4

13B

3.72 10-1

1119

200-510

8

0.76

0.84

40.2

3.5

0.8

81.6 1.3

86.5

5.5

-2.1

13C

2.91 10-1

1136

250-510

7

0.72

0.84

28.1

1.6

0.5

85.0 1.8

87.7

6.2

-5.4

13D

1.71 10

-1

1041

200-430

6

0.56

0.78

17.5

4.8

1.9

76.9 1.9

77.7

5.7

-3.4

14A

4.22 10-2

59

100-540

13

0.86

0.90

36.3

3.4

1.1

52.9 1.1

70.9

1.3

-2.0

15

14C

1.67 10-1

150

150-510

9

0.74

0.85

22.2

2.4

3.5

60.6 1.7

85.9

3.6

-0.8

14D

1.74 10-2

33

100-430

8

0.48

0.84

16.2

3.5

2.7

65.9  1.6

81.6

3.8

-1.0

Table S3: Archaeointensity results for the characteristic components at specimen level. Site, name of the archaeological site; Specimen, laboratory code for each specimen; NRM, intensity of the Natural Remanent Magnetization in A/m ; χ, initial susceptibility in 10-6 SI units (for the ceramic samples NRM and χ are normalized by the weight); Tmin-Tmax, interval temperature used for the slope calculation; n, number of data points within this temperature interval; f, fraction of the NRM component used for the slope calculation; g, gap factor; q, quality factor; MAD, maximum angle of deviation; DANG, deviation angle; F ± sd, intensity and standard deviation per specimen without TRM anisotropy correction; Fe, intensity per specimen with correction of TRM anisotropy; ∆TRM, cooling rate correction factor per specimen; alt., alteration factor occurring during the cooling rate protocol. Note that the laboratory field (between 40 and 60 µT) was applied along the z axis of the specimens. The type of material and the directional archaeomagnetic results available for the four studied kilns are also indicated.

16

Name

TM-02

Age

Microwave energy

(yrs AD)

(J)

800-850

19-81

n

7

f

g

0.53

0.81

q

MAD

F  sd

Fpo

Differences

(°)

(µT)

(µT)

(%)

1.5

78.7 ± 1.2

68.8

12.6

1.0

75.2 ± 1.1

-

0.3

83.7 ± 0.6

80.6

3.7

28.0

TM-03

too weak magnetization

TM-04

rejected: factor f lower than 0.50

TM-05

750-800

29-250

18

TM-06 TM-07

0.88

0.90

57.0

too weak magnetization 750-800

44-230

15

TM-08

0.82

0.92

111.0

too weak magnetization

TM-09

rejected: failed pTMRM checks

TM-10

significant secondary component of remanence

TM-11

significant secondary component of remanence

TM-12

rejected: failed pTMRM checks

TM-13

775-800

30-124

7

0.81

0.78

29.0

0.7

114.7 ± 2.5

81.7

28.8

TM-14

850-900

30-90

7

0.54

0.81

21.0

1.4

81.5 ± 1.7

78.7

3.4

Table S4: Microwave archaeointensity results obtained for the ceramic fragments. Name, name of the specimens; Age, age of the corresponding fragments; Microwave energy, interval of microwave energy used for the slope calculation; n, number of data points within this temperature interval; f, fraction of the NRM component used in the slope calculation; g, gap factor; q, quality factor; MAD, maximum angle of deviation; F ± sd, intensity and standard deviation per specimen; Fpo, TRM anisotropy corrected weighted mean intensity per fragment calculated from Thellier experiments; Differences, differences between microwave results and mean fragment Thellier values.

17

5. Supplementary Figure

Figure S1: Microwave archaeointensity results. Examples of NRM-TMRM plots for representative samples derived from ceramic fragments. pTMRM checks (solid line arrow) and pTMRM tail checks (dashed line with white square) are shown. The corresponding microwave demagnetisation orthogonal vector projections are also showed. Open (solid) circles are projections upon vertical (horizontal) planes.

18

Improving our knowledge of rapid geomagnetic field intensity changes ...

Improving our knowledge of rapid geomagnetic field intensity changes ...

Suggest Documents

Evidence for rapid geomagnetic field intensity variations in Western ...

Geomagnetic Field Intensity and Suppression of ... - Google Sites

Low intensity of the geomagnetic field in Early Jurassic time

Wavelet Analysis of Geomagnetic Field Intensity for the ... - Google Sites

Changes in the Geomagnetic Total Intensity Observed before the

Changes in the Geomagnetic Total Intensity Observed before the ...

New archaeointensity data from Italy and geomagnetic field intensity ...

Changes of Geomagnetic Total Field and Occurrences of ... - CiteSeerX

Regional cosmic ray induced ionization and geomagnetic field changes

EDIFICE- Changes in the geomagnetic dipole (Earth Dipole Field

Global changes in intensity of the Earth's magnetic field ... - Aldebaran.cz

The geomagnetic field in Croatia

What is the Geomagnetic Field?!

Improving our knowledge of male mosquito biology in ...

Rapid changes of magnetic Field polarity in the late ...

Mechanisms of Geomagnetic Field Influence on

Analysis of 11 Myr of geomagnetic intensity ... - Wiley Online Library

Analysis of 11 Myr of geomagnetic intensity variation

The effect of cosmic ray intensity variations and geomagnetic ...

Centennial to millennial geomagnetic field variations

Ninth generation international geomagnetic reference field released

Using Geomagnetic Field for Indoor Positioning

Reduced geomagnetic field may affect positive

Geomagnetic field modulates artificial static ... - Semantic Scholar