European Archaeomagnetism: Progress and Problems
Much progress has been made since the seminal work of Giuseppe Folgheraiter (1856-1913) in the late 19th century. So much so that recent advances now make it possible to draw up complete isogonic and isoclinic maps for Europe and adjacent areas spanning the last three millennia (Pavon-Carrasco et al., 2009). Results based on multiple independent studies, with high precision and good age control are crucial and should be recognized as "anchor points" (e.g. Pompeii). On the other hand, the nagging problem of outliers persists. Among the possible causes are magnetic refraction, physical distortion, and inadequate chronological control. Some examples, drawn from our own investigations over the last 30 years, will be discussed in detail. These include previously unpublished data from a detailed study (more than 100 samples) of a kiln in southern Italy, and an apparently good (but aberrant) archaeodirection from a kiln in southern Spain.
Regional Archeomagnetic Model for Europe for the Last 3000 Years: Application to Dating.
Recently a new regional archeomagnetic model in Europe for the last three millennia has been proposed. This model, SCHA.DIF.3K (Pavón-Carrasco et al., 2009, Geochem. Geophys. Geosyst., doi:10.1029/2008GC002244, in press), is based on a Spherical Cap Harmonic Analysis (SCHA) for spatial representation and sliding windows method in time. The model provides information of both directional and intensity variation of the Earth's Magnetic Field for the last 3000 years in the European region. One of the immediate applications of SCHA.DIF.3K regional model is its use as tool for archeomagnetic dating. So far the PalaeoSecular Variation Curve (PSVC) determined for a region has been used for archeomagnetic dating. The limitation of this application is the distance from the dating point to the location of the reference curve (the relocation error). In addition it must be borne in mind that the PSVC are individually generated for each region, so there is no consistency enforced between curves from neighboring areas. The use of the SCHA.DIF.3K model as a tool for archeomagnetic dating represents an improvement for several reasons. First of all, the regional model has been generated considering all elements of the geomagnetic field (declination, inclination and intensity). Second, the regional model is built with an in situ archeomagnetic database. Furthermore, the database covers the whole time period from 1000 BC to 1900 AD, while the database used in the PSVC has gaps of data for any time interval. Finally, and more important, we can generate a PSVC at the location of the archeological structure, avoiding in this way the relocation error associated with traditional PSVC. To demonstrate the utility of the regional SCHA.DIF.3K model, we have used it to date several archeological structures and we have compared results with the archaeological information and/or archeomagnetic dating provide by the use of the PSVC.
The Mesoamerica Secular Variation Curve. A continuous research since 1999.
Since 1999 we have been working on improving the Mesoamerica Secular Variation Curve. We re-sampled some sites that Wolfman initially used in its first proposal of this Curve, such as Teotihuacan, Teopancazco and Tula. Wolfman 's curve only had 4 radiocarbon dates directly associated with the sampled sites; the other dating were actually stratigraphic and ceramic correlations. More than 28 radiocarbon dates have been incorporated from AD 60 to 560 from samples associated to Teotihuacan civilization, 13 more from Xochicalco and 10 from Tula. More than 1000 specimens, from 12 twelve sites were fully processed with alternated field demagnetization We have been working during the excavation campaigns and training the archeologist to get their samples. A 100-year moving window was employed to get the average poles. A Bayesian statistic has been employed in order to improve our curve. We still denote a lack of data from two time intervals: from 0 to 200 and from 1000 to 1600. We are now working on it, collecting samples from these periods, such as those from La Joya, Ver, which are now been processing and from which some preliminary results will be reported.
Dating Post-Medieval Archaeology: Which Global Geomagnetic Field Model to use?
The scientific dating of Post-Medieval archaeology (16th Century onwards) is problematic as most methods cannot provide any better resolution than may be apparent from contextual or stylistic considerations. As high resolution global geomagnetic field models exist for this period, archaeomagnetism offers the possibility of bi-decadal dating of burnt in situ structures, with implications for the management of cultural heritage. The question arises as to which global geomagnetic field model is most appropriate for this dating? Should the high resolution historical field model, gufm (Jackson et al., 2000, Four centuries of geomagnetic secular variation from historical records, Phil. Trans. Roy. Soc. Lond. A, 358, 957- 90.) which covers the period 1590-1990 AD and is based on data from ship's logs be used, or should an archaeomagnetic model such as GMADE2K.2 (Lodge & Holme, 2008, Developing a global geomagnetic field model for archaeomagnetic dating in Europe for the last 2000 years (updating GMADE2K.1 to GMADE2K.2), Geophys. Res. Abstr., 10, Abstract EGU2008-A-03470) be used? In general a higher accuracy can be expected from the historical model, but the modeling strategy for gufm is aimed at investigating the magnetic field evolution at the core-mantle boundary, whilst GMADE2K.2 is developed to serve as an archaeomagnetic dating tool. If we compare secular variation curves in Europe for declination at this time, then the two models agree very well. For inclination however, there is a discrepancy pre-1800 AD between the two models, with the historical model tending to higher inclinations. Here we study the possible causes of this discrepancy: How reliable are the early historical inclination data? How reliable is the historical model at this time - is the inclination being affected by the domination of declination data? Finally, are the archaeomagnetic data systematically low, possibly caused by undetected magnetic refraction? The advantage of constructing global geomagnetic field models is that the inter-dependence of the components is taken into account. However, if we cannot reconcile the archaeomagnetic models and data with the historical model and data, then dating remains problematic.
The Donegal Sign Tree: A Local Legend Confirmed with Holographic Radar and 3-D Magnetics
A tree at a crossroad in Historic Donegal, PA (founded 1722) bears unusual burls. Two are similar in size, and lie on opposite sides of the trunk at a height of six feet. Locals say that the tree engulfed an old road sign, and the geometry of the burls gives this appearance. However, the trunk between these two burls bears no welt where it sealed after swallowing the sign. In addition, there are other burls farther up the tree, which are not consistent with engulfed signs. Although the locals all know the legend of the swallowed sign, none ever actually saw the sign; not even an octogenarian who has lived at the crossroad his entire life, and recalls the tree as a child just as it is today. In order to test the veracity of the legend, this study performed subsurface imaging of the tree using holographic subsurface radar (Rascan), and 3-D measurements of the magnetic field about the tree using cesium vapor sensors. The Rascan system used is a continuous wave subsurface radar that operates at 5 discrete frequencies between 1.5 and 2.0 GHz. Reflections from subsurface objects are recorded as the phase difference pattern between an internal reference signal, and the reflected signal. Thus, it is a microwave analogy for optical holography. Rascan records reflections with two receiving antennae - parallel and perpendicular to the transmitter - so a single set of scans provides ten images; five frequencies at two polarizations. This ensures that an object at arbitrary depth will produce a strong phase difference in one of the images. As a consequence, elongate objects that are angled from the plane of scanning (e.g. a dipping sheet) produce "zebra stripes" of contrast values that vary cyclically with depth. The presence of stripes, and their relative positions in the different frequency images (the movement of which has been dubbed the "zebra shift") is useful for determining the relative depth of different portions of a dipping planar, or curved subsurface object. Rascan images of the tree revealed a reflector that produces a zebra shift pattern reminiscent of a curved reflector. However, given the curvature of the tree trunk, the zebra shift is more likely to represent a flat reflector beneath a curved scanning surface - consistent with the presence of the sign. As an independent confirmatory method, the tree was also subjected to a magnetic survey. First, the tree was swept with a magnetic locator - which indicated a magnetic target within the tree. In order to determine the configuration of this target, magnetic total field measurements were collected at the nodes of a 3-D grid surrounding the tree. The geometry of this survey is quite different from traditional archaeological prospection magnetometer surveys and, despite the relatively high latitude of Donegal PA, the vertical orientation of the suspected target mimics the common difficulties with magnetic surveys at low magnetic latitude. Therefore, the analytic signal was calculated to provide an easily interpreted magnetic anomaly that, together with the Rascan images, suggests that the story of the swallowed Donegal road sign may be true.
3-D Modelling of Magnetic Data from an Archaeological Site in Northwestern Tlaxcala State, Mexico
In Archaeology, geophysical methods had been applied usually in a qualitative form, limited only to the use of filters that enhance the data display. The main objective in this work is the implementation of a modeling technique that allows us to reconstruct the geometry of buried bodies and the determination of their depths. This is done by means of the estimation of the magnetic moments of archaeological objects using a three- dimensional mesh of individual magnetic dipoles using the least squares method and the singular value decomposition of a weighted matrix to solve the linear problem. The distribution and shape of the underlying archaeological remains can be inferred. This methodology was applied to an archaeological site called Los Teteles de Ocotitla, in the state of Tlaxcala, Mexico. A high-resolution magnetic prospection was carried out in three selected areas (terraces). The most important total field anomalies found on each area were inverted, obtaining results that were corroborated by archaeological excavations. This investigation demonstrates the potential of quantitative geophysical methods for the characterization of archaeological structures, in extension and in depth.
Archaeometric Prospection Using Electrical Survey Predictive Deconvolution (ESPD)
Once upon a time archaeological prospection was carried out mainly using electrical techniques. These days
magnetic techniques and GPR are used by preference. However, we have shown that electrical surveying
combined with the technique of predictive deconvolution is very effective at finding buried features where the
shape of the feature can be predicted in advance.
One such type of feature is the Grubenhaus (or sunken-featured, sunken-floored building, or SFB).
Grubenhaüser exist in the archaeological record as individual well-defined oblong pits that have been filled
and buried with other material. Aerial photographs at New Bewick in Northumberland, northern England (UK
Grid reference NU061206) showed quasi-rectangular features similar to those on aerial photographs at the
nearby Anglo-Saxon palace of Milfield (NT941339) which had been confirmed by excavation to be
Several electrical resistivity surveys were carried out over the area with an ABEM Mk II Terrameter and a
multiplexing box serving 31 electrodes in line at any given time. Both double-dipole and Wenner configurations
were used with an electrode spacing of 1 m. Data was acquired in blocks of 30 m by 30 m during a period of
dry summer weather while the field was under young winter wheat. The Wenner array produces a characteristic
'M' or 'W' shaped response over filled in excavations such as those expected to represent a Grubenhaus.
While this seems a disadvantage in the first instance, it can be used to improve the data. Such anomalies
were present in the raw New Bewick data. The resulting data were analysed using 1D and 2D predictive
deconvolution in order to remove the Wenner response. The deconvolution was carried out using an inverse
matrix element method.
The filtered results indicated the presence of an anomaly that is consistent with a Grubenhaus measuring
about 5 m by 4 m and with a pit depth of 0.6 m below 0.5 m of topsoil. The results also showed broader areas
of increased resistivity which have been attributed to compaction resulting from human and animal movement.
Following the geophysical study the site was excavated (T. Gates and C. O'Brien "Cropmarks at Milfield and
New Bewick and the Recognition of Grubenhaüser in Northumberland." Archaeologia Aeliana 5th series, Vol
XVI, 1988, 1-9) and a Grubenhaus was discovered at the site. The excavated Grubenhaus measured 4.7 m by
3.9 m with a pit depth of 0.5 m below the base of the topsoil. The deconvolved Wenner data performed better
than the double-dipole resistivity survey but was marginally slower.