New methods for investigating slag heaps: Integrating geoprospection, excavation and quantitative methods at Meroe, Sudan
Jane Humphris
UCL Qatar, HBKU, Doha, Qatar 25256, UK
Chris Carey
School of Environment and Technology, Cockcroft Building, University of Brighton, UK
Journal of Archaeological Science , Volume 70, June 2016, Pages 132-144
Outline
Highlights
Abstract
Keywords
1. Introduction: iron production in the Kingdom of Kush
2. Investigating slag heaps
3. Archaeometallurgical aims
4. Justification of methods
5. Materials and methods
5.1. GPS and total station survey
5.2. Gradiometer survey
5.3. Electrical resistivity survey
5.4. Excavation and quantitative sampling
5.5. GIS modelling
5.6. Mass and density estimates
6. Results
6.1. Geoprospection data
6.2. Volumetric data modelling of the slag-heap
6.3. Quantitative sampling
6.4. Integrating the data
7. Discussion
8. Conclusion
Acknowledgements
References
Abstract
This paper describes a multifaceted approach to the investigation of iron slag heaps, focusing on one of the slag heaps at the Royal City of Meroe in Sudan. This study marries together geoprospection data (gradiometry and electrical resistivity transects), topographic data and quantitative excavation data, to provide an analysis and comparison of the total volume, slag component and slag composition of a slag-heap. Significantly, the results demonstrate the limitations of using a topographic only model, but also demonstrate how volumetric modelling must be integrated within quantitative characterisation of slag-heap composition. In this case, quantitative sampling of the slag deposits revealed the composition of the slag assemblage was dominated by a newly defined category of slag which has major implications for reconstructing iron technologies in the Meroitic civilisation. This research highlights the dangers of applying simplistic models and basic investigative strategies to iron slag heaps and furthers the debate on applying volumetric modelling and excavation sampling to unexcavated areas of the finite and important resource of archaeometallurgical deposit sequences.
Keywords:
Iron slag, Sudan, Geoprospection, Quantitative sampling ,Volumetric modelling ,Meroitic civilisation
7. Discussion
The analyses of these data sets have demonstrated the applicability of marrying together geoprospection data, excavation data and quantitative sampling to produce volume estimates of entire slag heaps. They confirm that the depth and form of the slag-heap, and the quantity of archaeometallurgical material within are highly variable. In this example, the gradiometer data allowed the spatial definition of the slag-heap based on magnetic signature, whilst the electrical resistivity modelling allowed the depths of the deposits to be estimated. The electrical resistivity model clearly identified different components within the slag-heap, which the excavated data demonstrated to be an earlier building. This was not evident from the surface survey and consequently, the volumetric model of the slag-heap produced through a surface survey model overestimated the amount of archaeometallurgical remains within the slag-heap.
However, without full excavation of the slag-heap, it is impossible to know which of these models is closest to reality: all have unavoidable limitations in data collection. Primarily, the resistivity model was constructed from three transects, with some of the quadripole measurement points unusable due to the sub-surface conditions. In addition, the electrical resistivity model interpolates between points, and interfaces are interpreted from this interpolation, all of which provide a degree of approximation of the position of the interface. The excavation data only focused on part of the slag-heap, so some areas (in fact the majority of the slag-heap) are unknown from an excavation perspective.
The quantitative sampling aimed to populate the volumetric models, as well as characterise the archaeometallurgical materials within the slag-heap. This data revealed complexity and variation within the deposit sequence of the slag-heap, a trait common throughout all slag heaps excavated so far at Meroe. As a result, archaeometallurgical deposits were shown to have a remarkably consistent percentage of total slag as a percentage of total archaeometallurgical material. This can be interpreted as a significant degree of standardisation of the smelting process, producing similar masses of archaeometallurgical waste per smelt and throughout the lifetime of the workshop(s) operation that produced the waste that formed MIS6. Such consistency hints at standardisation and routine of this industry.
The sheer volume and relative abundance of category 5 material comes out as a strong trend in the data (and is also noted during the excavation and slag processing of other slag heaps excavated during this research at Meroe). This category 5 slag is the dominant material within the archaeometallurgical remains of MIS6 and consequently is the most important in understanding the ironworking practices at Meroe. The small nature of the slag in category 5 is produced through the ironworking technologies employed at Meroe. Indeed, given the current state of knowledge of the smelting processes at Meroe and the dominance of this type 5 material, it calls into question whether this iron smelting should be considered as a classic slag-tapping industry. Further analysis of this important material group is underway and will provide a critical insight into the ironworking technologies at the royal city. The importance of category 5 slag has only been realised through the quantitative approach that has been detailed here. Although this is a time-consuming activity, the results support the effort to continue with this slag processing strategy to provide a more complete understanding of the excavated deposits.
It is also a sage point to note that often sample collection from slag heaps will focus on larger, more easily understood pieces of slag, e.g. tap slag, and will ignore smaller pieces such as category 5 slag. Such unrepresentative sample selection can become the basis of laboratory investigation, and subsequent reconstruction of past technologies. It is clear that at Meroe at least, such approaches would be unrepresentative of the slag assemblage as a whole. However, it is important to realise that different projects have different scales, experience, resources and field conditions for sampling of archaeometallurgical remains, and that smelting techniques across space and time are variable and so leave variable remains. The issue of sampling is important for process understanding, and has been previously inadvertently ignored in some field programmes, hence the reason for critical discussion of this issue in the archaeological literature (see earlier section) and this paper. Across the sub-discipline of archaeometallurgy even when quantitative sampling has occurred, it is usually not to pre-defined standards that have been robustly and statistically tested against other sampling methods. There is not a body of literature that the authors are aware of that discusses the excavation of entire slag heaps, analysing all samples and retrogressively modelling suitable representative sampling methods for that particular site. Consequently, this is a major gap in the field which requires synthesis and critical analysis of data from academic and commercial spheres.
The quantitative sampling also revealed insights into depositional process. Variability was discovered in the composition of the slag-heap between trenches. Trench 3 had the highest amount of archaeometallurgical material, the highest mass and the highest density of total slag, and was excavated in the most slag-dense area of the slag-heap. Trench 2 demonstrated the lowest density of archaeometallurgical materials. The reasons for this difference at trench 2 are unclear, but one possible interpretation is the deposits in trench 2 represent a deliberate infilling episode of part of the earlier building by human agency, through the re-deposition of slag. A further intriguing aspect of trench 2 is that a number of kg of pottery were retrieved, which was far higher than trenches 3 and 4, indicating a different depositional history to trenches 3 and 4. Further analysis of this trench 2 material is ongoing.
The quantitative data from trench 4 is easier to interpret, with a density of archaeometallurgical materials that was between the values obtained from trenches 3 and 2. On excavation this trench was interpreted as having hiatuses between some slag dumping episodes, with contexts evident in section containing sand/sediment, deposited between contexts rich in archaeometallurgical materials. The presence of sand dominated contexts between archaeometallurgical waste contexts indicates some form of periodicity in the deposition of iron slags, at least in trench 4, i.e. multiple smelting wastes deposited in the same place before a hiatus/sand accumulation, before a further campaign of smelting.
8. Conclusion
On a pre-excavation basis, topographic modelling, electrical resistivity survey and gradiometer survey have been demonstrated to be a powerful package for the estimation of archaeometallurgical structures and deposit sequences within slag heaps. This has allowed the positioning of targeted excavation areas and key facets of slag heaps to be identified before excavation, such as the underlying buildings and depth of archaeological deposits at MIS6, guiding the positioning of the excavation trenches to answer specific questions.
The quantification of slag categories from the excavations has provided significant insight, which can be coupled to both the laboratory and geoprospection data sets. Critically, the importance of the category 5 slag has been brought into focus; a defining characteristic of the slag assemblage at MIS6. The quantification process has demonstrated that selecting a representative sample of ironworking deposits, such as those at Meroe, requires a detailed analysis of the structure of the slag heaps: surface collection of individual pieces or hand picking during excavation is liable to yield an unrepresentative sample population for further analysis.
The combination of the quantitative data with the geoprospection data has allowed the volume and mass of the archaeometallurgical materials within MIS6 to be estimated. However, these calculations demonstrate just how difficult it is to estimate slag volumes and deriving production figures. The data modelling clearly shows that topographic survey alone is insufficient, and although the combination of methods used in this study represents a step forward, new emphasis needs to be given to geoprospection methods within ironworking site-scapes, combined with detailed quantitative sampling.
The number of large-scale excavations in academic research has arguably diminished in recent decades, with archaeologists well aware of the old archaeological maxim – ‘excavation is destruction’. Comparatively smaller excavations are now more common, seeing increasingly complex methods of scientific analysis to increase data yields from smaller interventions. Due to financial constraints as well as more rounded academic appreciation of the finite nature archaeological record, smaller excavations are being used more and more to construct models of reality. More data is needed and expected from these smaller excavations to make our interpretations closer to the archaeological realities. In contrast, the opposite maybe true of western commercial field projects driven largely by development led archaeology where largescale excavation of slag heaps is more frequent, but this is not feasible at sites of such importance as Meroe. Indeed the most recent Historic England Archaeometallurgy Guidelines for Best Practice, are written mainly for, ‘curators and contractors within archaeology in the UK’ (2015, p. 1), highlighting the different challenges facing archaeometallurgists working in different sectors of the discipline. In the academic sub-discipline of archaeometallurgy it is usually not possible or appropriate to completely excavate such large slag heaps (Crew, 2002: p, 165, 180). Therefore, finer models of characterisation are required to maximise knowledge whilst minimising impacts. Much work is yet to be done at Meroe, both in the laboratory and the field. However, by attempting to provide an integrated programme of geoprospection, excavation, quantification and the ongoing laboratory analysis, the secrets of the Meroitic slag heaps are slowly being revealed.
للقراءة والتحميل اضغط هنا
ليست هناك تعليقات:
إرسال تعليق