Abstracts
NAVIS ONE: Challenges and Opportunities for RSE and RDM by designing a common web application from a 1990s online database
Florian Thiery/Dr. Allard Mees (Römisch Germanisches Zentralmuseum, Leibniz Research Institute for Archaeology, Department of Scientific IT, Mainz, Germany)
Although the current situation in archaeology still shows a clear predominance of relational databases, the relevance of LOD and the Semantic Web in the Digital Humanities has increased substantially. Enhancing SQL databases by applying LOD technologies enables the application of the RDF standard, semantic modelling, interlinking of data, and Open Access by offering SPARQL endpoints to create FAIR Linked Open Usable Data. Creating a new infrastructure, an adapted data model, and the implementation of interactive web applications based on legacy systems causes challenges for Research Software Engineering (RSE) and Research Data Management (RDM) issues. This talk will demonstrate the example of NAVIS ONE.
The NAVIS Ship Databases 1-3 at the RGZM cover ancient ships (< AD 1200). The European Union funded the NAVIS projects, and its research community consisted of scientists from more than ten countries. NAVIS I (1996-1999) comprises shipwrecks, reconstructions and models, NAVIS II (1999-2001) focuses on ancient ship depictions on objects, and NAVIS III (2002) is related to ships on Roman coins. Moreover, the NAVIS ship databases enclose terms organised in controlled vocabularies, i.e. NAVIS I: thesauri of ship components and wreck-specific thesauri; NAVIS II: thesauri of ship components and their properties, objects keywords, dating, material as well as techniques; NAVIS III: thesauri of emperors, mints, coin types and ship features.
Designing a common web app causes challenges in merging three different individual MS Access data models focussing on wrecks, models, reconstructions and depictions of archaeological objects and artefacts, especially coins, in a common data model based on PostgreSQL. Furthermore, transforming, cleaning, validating and normalising data is a highly elaborate task. These refined data enable opportunities in Semantic Web applications, LOD
and community work, e.g. NFDI4Objects TRAILs. Publishing data in the LOD hub archaeology.link connects them within the LOD Cloud, e.g. by interlinking ARS in NAVIS ONE and in ARS3D as well as iDAI.objects and entities of the Wikidata Knowledge Graph; thesauri can be published via a hub for authority files and community-driven vocabularies: DANTE; Nomisma can be used to share coins as a domain-specific hub.
3DWorkSpace – an open science/interactive tool for 3D datasets
Tijm Lanjouw/Jitte Waagen (University of Amsterdam, 4DResearchLab)
The continued growth of digital methods and practices within the humanities over the past 30 years has provided a wealth of datasets that are transforming the discipline at an astonishing rate. New tools are urgently needed for 3D datasets to improve accessibility, facilitate engagement and interaction with the datasets, and promote two-directional knowledge transfer. The 3DWorkSpace project, funded by the NWO Open Science Fund, aims to deliver such a tool by adapting the open-source Voyager 3D digital museum curation tool suite and promote interactive engagement with traditionally complex digital datasets. Embedded structured guidance, or Learning Pathways, will be generated to provide training for gaining competence and skills for interpreting 3D datasets. Through the creation of annotated personal 3D collections that can be tailored to specific learning goals or interests, broader narratives can be generated and new avenues for knowledge publication opened up. This presentation will report on the progress so far and gained insights.
GIS - painlessly. Increasing the efficiency of archaeological documentation through simplification of geodata management and geoprocessing
Gunhilt Merker/Jana Nolle/Szymon Domagała (Archäologische Ausgrabungen und Bauprojektbetreuung [AAB])
During our few-years-lasting adventure with GIS in German and Polish academic and investment archeology, we faced various documentation systems and standards for spatial documentation of archaeological features and artifacts. Enriched with experience from past projects, we present a series of good practices at various stages of building a spatial database, aimed at facilitating the processing and analysis of data from excavations using open-source QGIS software, based on examples of various projects undertaken across Germany in 2021-22.
Indigenous knowledge and Archaeoinformatics: Tapping into Indigenous Knowledge to advance computational models of hunter-gatherer societies
Eleftheria Paliou/Oliver Vogels/Lucía Cobo-Sánchez/Tilman Lenssen-Erz (University of Cologne)
Computational models of hunter-gatherer mobility and behaviour have long been used in archaeology to offer insights into the choice of site location, the subsistence strategies, survival resilience and social development of early human societies. Central to these models are foraging practices that would have shaped prehistoric human behaviour, human -animal and human-environment interactions. To date, ethnographic research has been instrumental in helping archaeologists theorise foraging behaviour, nonetheless several aspects of traditional hunting practices (i.e. hunting on foot, without firearms, or optical aids) have yet to be systematically explored and are still little understood. The project “Indigenous Knowledge and Archaeoinformatics” aims to bridge this gap by bringing together researchers from the University of Cologne and indigenous hunting experts. It offers a rare opportunity to capture using geospatial technologies the movements of Ju/’hoansi and Hai//kom San game tracking experts during hunting bouts in Doro!nawas, a mountainous landscape in western Central Namibia which is rich in prehistoric rock art. During three field seasons, which took place between 2019 and 2021, a variety of georeferenced quantitative data (including movement, caloric expenditure, wind speed and direction measurements), as well as qualitative information on decision-making during hunting were collected. This paper will discuss how such data can be used to inform and develop further computational models of past hunter-gatherer mobility and behaviour, and in particular GIS-based and agent-based models. More specifically, the focus will be upon qualitative observations and quantitative data analyses that offer new insights into wayfinding behaviour, human-animal and human-environment interactions in arid and semi-arid landscapes, and have the potential to better support archaeological model building efforts.
Timing and length of vegetation periods as a potential driver for long-term trends in the movement and distribution of hunter-gatherer populations
Louise Tharandt/Vlad Krakov/Florian Linsel/Patrick Ludwig/Andreas Maier (University of Cologne/Simon Fraser University/Martin-Luther-Universität Halle-Wittenberg/Karlsruhe Institute of Technology)
Migrating animals often follow the greening of the landscape during the onset of the vegetation period. Such movements likely have been of interest for prehistoric hunter-gatherers and thus may have affected long-distance migrations as well as seasonal land-use patterns and location of foraging areas. Timing and length of the vegetation period may thus have been an important driver in long-term trends of spatial patterning of populations. Consequently, shifting patterns of the greening and productivity of the landscape might reveal ecological pull factors with different directionality that might hold explanatory potential for the patterns we observe in the archaeological record.
In this paper, we present a protocol for estimating the timing and length of the vegetation period and explore its explanatory potential in two case studies, comparing differences between stadial and interstadial conditions. Our method consists of estimating the growing season by calculating the vegetation period, which was automated by using the programming language R, and producing images showing vegetation patterns with Python and QGIS. This means that our method and the results are reproducible. First, we look at a dynamic phase of population expansion related to the arrival of anatomically modern humans in Europe (i.e., the Aurignacian between 43 and 33 ka) and discuss to what extent our results match the archaeological observable distribution of sites and the hypotheses of migration corridors, such as the ‘Danube corridor’. Second, we explore the spatial patterning of hunter-gatherer populations in the topographically homogeneous landscape of the East European Plain between 40 and 27 ka. Despite the uniformitarian landscape, a comparatively large number of taxonomic units is reported from this region. Here, we evaluate to what extent the observed phenological gradients may have fostered a segregation of population in certain regions and thus the development of regional differences in the material culture.
Based on our results, we conclude that differences in the estimated timing of the vegetation period in different parts of the East European Plain as well as Western Europe are indeed helpful to understand the large-scale structure of the archaeological record in these areas.
NASSA’s open library of archaeological ABM modules: overview and examples
Andreas Angourakis/Dries Daems/ Philip Verhagen (Ruhr-Universität Bochum/Middle East Technical University & KU Leuven/Vrije Universiteit Amsterdam)
The primary mission of the Network for Agent-based modelling of Socio-ecological Systems in Archaeology (NASSA) is to create and maintain an open library of agent-based modelling (ABM) components. The initiative distinguishes itself in goal and scope from other open science libraries, such as the CoMSES Network Model Library. First, the library is designed to be a collection of modules, not models, formatted and documented independently following a common but flexible standard. Its overall aim is to facilitate its future re-use within the broader community of researchers. Second, the library is first-and-foremost archaeological, meaning that it is both pragmatically specialised, allowing more immediate use in our field, and intrinsically interdisciplinary, thus including modules on many subjects and domains.
This paper will briefly present an overview of NASSA and the library developed so far. Given the still early stage of the library, the paper focuses on giving enough instructions and examples to encourage future submissions. We will review the main aspects of NASSA organisation and infrastructure and offer a scheme of the library concept, including our definition of modules as either algorithms or sub-models. We will describe the submission process and, most importantly, explain the metadata and minimum file structure required for submissions. Next, we will offer three practical examples of the process of creating modules out of pre-existing models and/or code and preparing the related metadata for submission.
Reconstruction of ancient ecologically favoured zones in Western Sudan
Jana Eger-Karberg/ Tim Karberg (University of Münster)
Quantitative methods play an increasing role in landscape and geo-archaeological research, especially in remote areas. At the peripheries of the Kushite and Nubian cultures, entanglement of intensive agriculture and extensive rangeland economy at the fringe of the desert often depended on only slight shifts of climatic and vegetation zones. In landscape archaeology, the application of geo-scientific and, especially, geo-statistical methods on historical and archaeological content-related questions show a great potential for the diachronic modelling of different land use and subsistence strategies. The climatic, vegetational, and economic fuzziness of the rangelands of the Kushite and
Nubian peripheries at the fringe of habitability makes this zone very sensitive to climatic and other landscape ecological fluctuations. Therefore, the risk-minimizing capacity of entangled economic and land use strategies, incorporating extensive and intensified agricultural and pastoral modes of production, can be exemplarily studied at these rangeland peripheries. While geo- and agricultural sciences provide dense datasets about recent developments, archaeology can contribute extrapolation of these data towards a long-term diachronic perspective, which is crucial when dealing with the question of long-term sustainability of different subsistence and land use strategies as well as different models and levels of statehood and socio-economic organization.
Exemplary data from the Sudanese region of Northern Kordofan encompasses several classes of archaeological records from different environments, representing focal points of social organization on the peripheries of late Antiquity and Medieval states and cultures. The identification of ecologically favoured zones from satellite data analysis (including multispectral vegetation and soil composition analysis as well as mapping ancient drainage systems nowadays covered by aeollic sediment from satellite based radar data) and geo-hydrological modelling, statistically contextualized with archaeological data, allows the diachronic extrapolation of landscape ecological models into the past, and the correlation of archaeologically documented peaks of human habitation density with general climatic and ecological circumstances.
Digging in Documents - Creating a Search Engine for Archaeological Literature using Text Mining
Alex Brandsen/Karsten Lambers/Suzan Verberne/Milco Wansleeben (Leiden University)
The archaeological domain creates huge amounts of text, from books and scholarly articles to unpublished fieldwork reports. Easy access to the information hidden in these texts is a substantial problem for the archaeological field.
Currently it is generally only possible to search through the metadata of these documents, through various search interfaces of repositories and publishers. However, these metadata are often limited and sometimes inconsistent, and don't capture the ‘by-catch opportunity’; i.e. a single Bronze Age find within a large Medieval excavation, which isn't mentioned in the metadata, would be impossible to find. Currently, we see that archaeologists often gather large amounts of documents and then manually search through them, one by one, which is a time-consuming and inaccurate process.
In our research, we have been building a search engine (called AGNES) that can provide access to all of the text, in all of the collected documents, allowing for more efficient and targeted research. Besides simple keyword search (also called 'full text search'), we also use Named Entity Recognition (NER) to correctly identify and distinguish between archaeological concepts, such as artefacts, time periods, places, etc. This further enhances the search experience by decreasing problems with synonymy and polysemy (concepts that can be described by many terms, and terms that can have multiple meanings, respectively).
Previously we have focused on using ‘classic’ machine learning such as Conditional Random Fields. Recently, Transfer Learning (a subfield of Deep Learning) has been introduced to the field of natural language processing, promising the same improvements it had on the field of computer vision (the "imageNet moment"). In this paper, we will present our results using Transfer Learning for archaeological NER in the Dutch language.
Besides the technical details of the Transfer Learning approach, we will also present an evaluation of the AGNES search system, by means of a case study on the prevalence of Early Medieval cremations, performed together with Leiden PhD Femke Lippok. While it was generally assumed that the burial practice changed almost completely from cremations to inhumations in the period from the Roman Period leading into the Early Middle Ages, our research shows that cremations are much more prevalent than commonly thought. We detected 30% more cremations than previously known by experts in the field, leading to a more nuanced view of the burial repertoire in this period.
Releasing the knowledge hidden in PDFs by using NLP and NER
Alphaeus Lien-Talks (University of York)
Heritage data is of exceptional importance, not only in understanding the past, but in recognising trends for the future. The English Historic High Street is no exception (Carmona, 2015). However, often this data is recorded in PDFs and so is difficult to share and access at scale.
Relevant data can be very diverse in its nature, including information related to lifestyle, migration, demographics, economics, environment, climate change and housing needs (Wrigley and Lambiri, 2014; Kim, 2011). As such, lessons learned from this urban environment can lead to aid the understanding of our ancestors on an international scale. The diversity of data and its sources creates limits on its accessibility and interoperability (Meyer et al., 2007). These issues are extended further by the scale and diversity of new data types generated from laser scanning, LIDAR, GIS and HBIM (Tait, 2016; Emmanuel and Loconsole, 2015; Dore and Murphy, 2012). These all contribute to the “Big Data” of the historic built environment (Gattiglia, 2015). One particular area of investigation is the approach to make data contained within PDF format more FAIR (Findable, Accessible, Interoperable and Reusable). To aid with this, collaborations are taking place between the universities of York and Leiden to develop technological capabilities of Natural Language Processing and Named Entity Recognition to release the data locked up in PDFs. Consequently, using the English Historic High Street as a case study, it is possible to aid the understanding of what must also be done for similar datasets within the Netherlands and Germany.
Classification of pottery assemblages in archaeology: a machine learning approach
Guilherme D'Andrea Curra (Leiden University)
Research aim and questions
Artefact classification is one of the main themes and an important practice since the beginnings of archaeology, while machine learning (ML) became one of the most efficient approaches to increase our knowledge in a number of disciplines. This research describes a ML model developed for the classification of pottery assemblages, identifying its benefits and limitations, focusing on the importance of artefacts features for the identification of vessel shape classes, the relations among these classes, and to what extent this kind of knowledge can be used to replicate classifications made by experts.
Data and methods
The dataset source is an assemblage of 496 pottery vessels representing nine shape classes and four archaeological sites from the Bronze Age Northeastern Syria (Upper Mesopotamia), made available by the Arcane (2016) project. The classification methodology was based on principles of quantitative archaeology, using vessel measurements and categorical features, implemented by supervised and unsupervised learning ML algorithms and supporting methods from the scikitlearn and SciPy libraries. The Anaconda platform, the Jupyter notebook environment and ImageJ for image processing complete the main software used in the research.
Results and interpretation
The research results indicate benefits and limitations in the application of ML models in the classification of pottery assemblages. The limitations are especially related to number of samples versus target classes, the homogeneity of the vessels context in the dataset, and the quality of data available for the samples. The overall results indicate the ML model can replicate the classifications made by experts with an F1-Score of at least 80% in 2⁄3 of the cases (Table 1). The results suggest that the model can be useful to experts, assisting in the identification of the most relevant artefact features (Figure 1) and similarities among classes of artefacts, as well possible misclassifications, ultimately providing new insights into the classification of pottery assemblages in archaeology.
Interlinking Big Exchange – Exceptionally large raw material networks in prehistoric Europe and beyond
J. Hilpert/S. Strohm/T. Kerig (Christian-Albrechts-Universität zu Kiel)
The focus of this paper is on far-reaching, interregional exchange networks of specific raw materials with a clearly determinable source. We present a collection of trans-Eurasian datasets (Neolithic to Bronze Age) that is being assembled within the “Big Exchange” projects, a collaboration between archaeology and data sciences at the Cluster of Excellence ROOTS and a data science/archaeology tandem project at the KI@CAU Datencampus. These projects additionally provide the platform for a moderated but basically open cooperation of interested parties, connecting archaeologists from all over Europe working on raw material networks. Data brought together by this working group is used as a case study of cross domain integration of data from different sources, with a focus on automation potential, domain requirements and upcoming challenges in the process. The underlying infrastructure and integration process has been developed by a subproject of the CRC 1266 “Scales of Transformation”.
Research into prehistoric exchange networks focused on the spatial distribution of single goods, especially very common or very rare raw materials. Important examples are on one hand side the different flint varieties (e.g. Rijckholt, Lousberg, Szentgál, Grand Pressigny, Monti Lessini) and common metals used for daily purposes and on the other hand exotic and rare materials like jade, ivory, lapis lazuli, kauri, spondylus or coral. These goods have always been exchanged, in the Neolithic on an interregional and continental, from the Bronze Age onwards also on intercontinental scales.
A joint consideration of different raw material distributions under a network perspective allows the recognition of temporal and spatial patterns that have not been recognised in this clarity so far. Thinking of exchange from a network perspective requires actors, which are linked for example through commonalities or interlinked activity in exchange processes. So far, studies focussing on single materials, predefined time frames or smaller regions have been able to rely on substantial expert knowledge of archaeological research in order to identify these actors. The ongoing data collection covers numerous materials, touches three continents and several millennia, thus including all the existing expert knowledge would be more than a lifetime’s task. A method combining Kernel Density Estimations (KDE) for materials present in a grid-based partition of the geographical space was adapted to enable time-independent, regionalised comparisons of network structures with the goal of identifying virtual actors as a foundation for network analysis and further analysis.
We will present examples reaching from Western and Central Europe to the Aegean as well as to Western Asia. Two case studies - from the early Neolithic (LBK) and the Aegean Middle Bronze Age – are presented. Through the application of network analytical tools, the formation of clique within and between regions can be observed. Possibilities and limits of the presented approaches will be discussed.
Impact of Levantine transit trade on the cities of Transylvania, Wallachia and Moldova in the 17th century
Daria Stefan (University of Ulm)
After losing Zadar, Sigismund of Luxemburg, and his ally the Republic of Genoa declared war on the Republic of Venice, which was the major spices supplier for Europe at the time. Consequently, an alternative supply chain developed: Levantine goods brought by the Genovese to the ports of the Danube and the Black Sea were transported by Wallachian and Transylvanian merchants to Buda, Kosice, and Lvov. By the beginning of the 16th century, five times more Levantine wares were reaching Buda by this route than the amount supplied through the Adriatic Coast.
This paper focuses on the impact of Levantine transit trade on the cities of Transylvania, Wallachia, and Moldova during the 17th century.
By modelling the trading routes of this area as a network, we simulated the flow of Levantine goods and identified two types of advantageous network positions: trading centres and potentially influent cities.
We characterized trading centres as cities with high Closeness and Degree Centrality scores, and defined influence as a combination of high Betweenness score, a low Constraint score and access to Structural Holes.
We also analysed the possibilities of brokerage between the three regions and how the network position may have affected a city’s effectiveness.
Statistical SNA provided the framework for quantifying the individual economic and political advantages of the cities in this area, as established through their involvement in Levantine trade.
Between Migration and Exchange: Addressing the Koriabo pottery and the late Carib-speaking expansions by means of Least-Cost-Path Analysis
Bruno de Souza Barreto (University of São Paulo / University of Bonn)
From AD 1200 to 1500, a pottery style called Koriabo appeared in most parts of Northern Amazonia, including Guianas and the Lower Amazon. Such ceramics were also present by the time of first colonial enterprises in the Caribbean, in sites dated from the 16th and 17th centuries in Lesser Antilles, which are understood as related to Kalinago peoples. Despite its distribution over a vast territory, the homogeneity of decorative forms and vessel shapes is a striking characteristic of such ceramics, whose most radiocarbon datings range from AD 1000 to 1500 in Amazonia. Some have raised that such ceramics would be understood as one of the archaeological correlates of the Carib-speaking peoples, in reason of its presence matched with places historically occupied by them. On the other hand, others agree that their phenomenon is more related to the formation of long-distance exchange networks. My hypothesis is that Koriabo could be explained by means of a two-folded process, including the making and reproduction of exchange networks and substantial population movements. Such migrations probably had a rapid rhythm and created more homogeneous patterns of archaeological visibility. One of the most accepted explanations for Carib-languages diaspora points out that the last expansion step initiated by AD 1000, when population growth and the strengthening of exchange networks led a diasporic movement to the regions known at the time of colonial encounters. The purpose of this paper is to discuss such ideas by the means of Least-Cost-Path Analysis and Time-Series modelling of radiocarbon Datings, to test how the archaeological data fits on the linguistic models for Carib languages expansions.
Testing an approach for close-range visibility analysis
Irmela Herzog (The Rhineland Commission for Archaeological Monuments and Sites, Bonn, Germany)
Since viewshed computation became readily available in GIS software packages several decades ago, visibility analysis has been included in quite a few archaeological studies. Nearly all of these studies focus on watching distant items such as signals of a watch tower. The study presented here was inspired by a site in Ecuador which might have been the location of events such as a religious ceremony about 500 years ago. For a given event location, the aim is to estimate the maximum number of spectators that could watch the event comfortably while sitting on the ground. The assessment of a spectator position does not only depend on the viewshed and the distance to the event. Additional factors to be considered are the slope of the spectator position, the angle of vertical and horizontal head movement and the probability of someone in front is blocking the view. Several publications suggest tolerable and comfortable ranges for each of these factors, and not all of them agree. A compromise was found, and a GIS procedure implemented that assesses spectator positions. But are these assessments correct? Is the resolution of the available elevation data adequate? This is tested by applying the GIS procedure to two sites in the Rhineland that serve or served as open air theatres. One of them is an amphitheatre that has been preserved since Roman times. Some changes were documented, but the theatre is still in use today and may accommodate an audience of more than 1000 people. The other site is larger, up to 2000 spectators attended events such as opera performances at this location. After opening in 1923, this so-called natural theatre was in operation for several years. Nowadays this area is hardly accessible anymore. ALS data is available for both sites as well as contour line data. The questions to be answered include: What is the impact of using contour lines instead of ALS data on the quality of the result? Does the GIS procedure correctly identify the known spectator locations? Are the parameters chosen correctly for assessing spectator positions? Are there any additional relevant factors that were not considered in the initial implementation of the GIS procedure?
Large-Scale aerial surveys in Mongolia: Direct versus automated georeferencing methods
Hendrik Rohland/Marco Block-Berlitz/Christina Franken (HTW Dresden/DAI Germany)
One of the most persistent challenges in archaeological documentation has always been to find a middle ground between technically advanced, often very expensive, time-consuming and the minimum necessary investigation methods that suffice to answer the archaeological questions at hand. With the use of multicopters in excavation documentation and for survey purposes, the question always arises: How good are the results obtained by automated aerial surveys? The scientific comparison of photogrammetrically produced 3D models with those produced by a laser scanner has a long tradition. The arguments for the use of photogrammetry are strong enough to find this technique in most documentation campaigns, especially where budget is an issue. But what about georeferencing? Multicopters, which are able to determine their position in space more or less exactly, provide so-called GEOTAGS to the camera images. Based on the images obtained in this way, 3D models can be georeferenced automatically. But how good are these results, for example, in comparison to RTK solutions?
We want to answer this and related questions with the results of the large-scale aerial campaigns in Mongolia's Orkhon Valley as presented in Block-Berlitz et al. in 2021 (see Fig. 1, for the data publication see Hüttel et al. 2021). While it is out of question that the highest accuracy possible should be strived for, where funding and available equipment allow for it, we want to show that there are indeed justifiable situations in which the biggest challenge, the cost factor, can be significantly reduced with an acceptable tradeoff in accuracy.
Application of various Ground-Penetrating Radar antenna systems and processing software on an archaeological test site
R. Kniess/A. Fediuk/H. Zöllner/R. Freibothe/N. Allroggen/B. Ullrich (Eastern Atlas GmbH, Berlin/University of Potsdam)
Ground-penetrating radar (GPR) has been used as a prospecting method for archaeological issues since the 1980s. Depending on antenna frequency and ground conditions, spatial resolution in the cm- dm range and depth penetration of about 1.5-3m can be achieved, as well as sufficient material contrasts of archaeological objects. In our study, we apply different GPR antenna systems on a test site with multiphase archaeological and modern structures at different depths. For this purpose, we use single and multi-channel systems as well as antenna arrays with and without step-frequency continuous wave technology. The antenna systems show different characteristics and potentials. Single-channel measurements can be used flexibly and without much effort. Multi-channel systems and antenna arrays allow faster areal measurements. The usage of the step-frequency technology allows to store data in frequency domain. This enables the user to reprocess data with different frequency weighting to enhance the features of interest. We first focus on challenges during field measurements. As an example, the handling of the antenna arrays must be carried out with an ATV due to their size/weight. Another point is the positioning, which we have carried out and compared in various measurements using marker wheel, total station and RTK-GNSS. In the second step, we will discuss the data processing. The major challenges here are how to deal with the large amount of data generated and how to visualize relevant 3d-structures in the ground. Processing and visualization of data might be quite different depending on the software used. We will show both manufacturer-independent PC and cloud applications, as well as manufacturer-provided software. An important aspect in multichannel and array measurements remains dealing with the not quite identical design of the different antennas and resulting stripes in depth slices. As the third step, we present the results of our test measurements. We provide radargrams and depth slices of the different antenna systems and processing software. In addition, we investigate the extent to which these results are comparable. From this, we draw conclusions on future prospections.
Hof van Maarland: underground, in the air, in bits and bytes
Joep Orbons (ArcheoPro)
To test prospection methods under different conditions, a site with known archaeology is needed. The site should have no planned development so the test site can be used continuously. One such site was found in the south of the Netherlands.
The 1.5 ha site holds a farmstead called “Hof van Maarland” dating back to the period 1500 to 1600. The farmstead was not stationary, it followed along with the medieval agricultural cultivation. After 1600 the village moved northwards and the site has since been abandoned and converted into an orchard, grassland and is now cropland. Because the site had a very short building history, it forms an ideal testing site.
The prospection started in March 2020 with EM, resistivity and magnetometry surveys and drone photographs. Over the next 2 years, parts were remeasured several times and over 30 drone photographic sets were made in ortho and oblique photographs.
The EM surveys show the location of the main building and the general layout of the farmstead. Resistivity gave more details about the buildings but is susceptible to dry soil conditions. A magnetic survey didn’t seem to give useful results in the first place. A more detailed survey will take place when the crops are harvested in the autumn.
The drone photographs produced some very useful research data and archaeological insights. After monitoring three different crops over a two year period, it shows that maize and clover did not present the archaeology but winter wheat did, under specific circumstances. During the slow growth in wet conditions, the wheat did not show the archaeology. Yet during two weeks of strong growth in a very dry period, the wheat shows discolorations over the known walls. Two weeks later, in wet conditions again, the discoloration disappeared. Due to the delay in growth, the length of the wheat over the walls never caught up. Drone photographs under specific grazing light produced some good archaeology during the further grow and ripening of the wheat.
A special situation occurred just before harvesting. The delayed growth of the wheat over the walls resulted in smaller length and in lighter grains, staying upright in the very latest phase of the ripping of the wheat. It resulted in erect wheat over the buried walls in contrast to flattened wheat everywhere else, showing the walls in a both visible as well as physical presentation.
The conclusion from the drone photography is that the discolorations due to growth-stress produces the best cropmarks but they are only visible during a very short time window. The results from the growth-stress will be visible for a much longer period of the crop season but the results are less clear.
The geophysical surveys and the drone photographs will be continued and any other prospection methods are welcomed to be tested on this site for years to come, hoping to gain more insight in different soilmarks, cropmark and instrument conditions.
Geophysical, Aerial and Thermal Prospection at Magoula Plataniotiki
Jitte Waagen (Amsterdam Centre for Ancient Studies and Archaeology / 4D Research Lab)
The Classical/Hellenistic settlement of (probably) Halos is located on an artificial hill called Magoula Plataniotiki close to the modern shoreline. The site has been surveyed through geophysical techniques in 2011 and 2015 using magnetic gradiometry, soil resistance, EMI methods (GEM2 and CMD), and GPR techniques (Noggin Plus 250MHz antennas). The geophysical survey was carried out over much of the northern half of Magoula Plataniotiki, covering about 40,000 square meters of built area. Google Earth and WorldView 2 satellite images of 2013 were processed in correlation to the geophysical images. In 2021 and 2022, the site was also investigated through drone aerial remote sensing methods. Visible light (Zenmuse X5S), multispectral (Micasense Rededge) and thermal infrared (Zenmuse XT2) sensors were employed. This covered the whole Magoula. As the potential success of aerial remote sensing is dependent on the seasonal conditions, e.g. of the crops, multiple aerial campaigns in summer, late autumn and early spring have been carried out. This maximizes the success rate of the aerial reconnaissance and allows for a comparative study of atmospheric/climatic variables such as diurnal temperature flux, ground humidity and ground vegetation.
The combined geophysical and aerial reconnaissance results provide a valuable composite dataset. Based on the GPR data we can conclude that the density of structures is high all over the surveyed area. Considering both survey seasons the settlement seems to expand to a maximum of 260m E-W. While GPR covered only about 135m N-S, the thermal and multispectral data indicate that building extends all the way to the south edge of the site, another 220m. Although expected this presence of structures in the previously uninvestigated southern part of the site was not confirmed before.
The added value of the drone survey lies mainly in its complementary nature to the geophysical investigations, in combination with the possibility to cover large areas in a single flight. The cumulative research data provide a wealth of information regarding the spatial layout of the site, suggesting a quite regular Classical/Hellenistic town plan with a dense clustering of orthogonal structures. The central and west-central regions of the site are occupied by multi-room rectilinear buildings with orientations slightly west of true north. Further to the east, GPR survey and satellite imagery indicated a diagonal northwest-southeast alignment, which excavation has shown to belong to an earlier (Classical) occupation phase.
Non-invasive approach to Ancient Pueblos: Results of geophysical surveys by Sand Canyon-Castle Rock Community Archaeological Project in Colorado, USA (2011-2019)
Piotr Szczepanik/Radosław Palonka (University of Cologne/Jagiellonian University Kraków)
The paper is focusing on non-invasive methods of prospection and subsequent analysis in various software of archaeological sites located in three canyons of southwestern Colorado in the North American Southwest. There are around forty Ancestral Pueblo culture sites with sandstone architecture and rock art placed in canyon alcoves or very close to them; also, there are older settlements located on open spaces as well as farming fields and possible water reservoirs. All sites are dated roughly from the 7th till 13th century A.D. This area is within the Canyons of the Ancients National Monument, the legally protected area managed by the US Bureau of Land Management. The project focuses on social and cultural changes in the area, during its inhabitation by Pueblo communities.
Native American populations were inhabiting this region from the pre-Columbian period and left an enormous amount of archaeological traces (Cordell 1997). Around the area of study, there are various sites that have different functions and diverse dating. Ancestral Pueblo groups created a micro-regional community of allied sites that were changing through the centuries. Using non-invasive methods of survey and documentation we would like to trace and line up the history of the inhabitants of this micro-region, the demographic fluctuations and migrations. Part of the project is devoted to geophysical surveys around some sites where geophysics were applicable, and during the seasons from 2011 to 2019 geophysical research was conducted in around ten archaeological sites. Also The common methods used there were electrical resistivity and magnetometry (with some addition of GPR) which gave us satisfying results revealing many of the unknown archaeological buildings. We would like to present some of them to show how much new information about this settlement community these methods provided for us, as well as showing their strengths and weaknesses on this specific landscape and archaeological sites.
Algorithmic Approaches for Visualising Heat-Induced Change in Burned Bones
Priscilla Ulguim/Tim Thompson (Teesside University)
Burned bones are frequently recovered from archaeological and forensic contexts. During combustion, heat-induced changes transform the chemical and physical microstructure of bone, resulting in dimensional and surface colour alteration dependent on environmental and burning conditions.
Heat-induced changes present taphonomic indicators which may be used to interpret information regarding the individuals, pre-burn and burning conditions, the aetiology of heat-induced changes, and funerary practices. Therefore, the quantification and documentation of patterns in heat-induced change in burned bone are vital for archaeological and forensic investigation, in particular for posterior and remote analyses, as well as where cremated remains are subject to reburial and repatriation. Furthermore, in the global South this documentation may perform a paramount role in future archaeological research in case of loss or destruction of remains resulting from disasters in heritage institutions.
However, the complexity of macroscopic changes leads to challenges in documentation, particularly in the differentiation and quantification of fracture types and surface colour change. Improved methods for quantifying and qualifying colour and morphological change are required which can objectively document variability while reducing handling time of fragmented and fragile material, and enabling more advanced, collaborative approaches such as crowdsourcing analysis.
The objective was to test the efficacy of algorithmically generated imaging in the recording of full surface morphology including fractures and colour change to support posterior analysis. Reflectance Transformation Imaging (RTI) was selected for this test as it allows users to view all possible lighting angles across surfaces by combining multiple images.
RTI has been applied to visualise and identify surface modification in bone artefacts from archaeological sites but has not been applied to assess heat-induced change in burned or cremated remains until the present work.
Future work aims to incorporate the results presented here to enable a more qualitative and quantitative method for burned bone surface modification by integrating Artificial Intelligence (AI), computer vision techniques, and convolutional neural networks.
Interactive visualisation of volumetric CT data in a museum context: the animal mummies of the Allard Pierson
Markus Stoffer/Tijm Lanjouw (4D Research Lab, University of Amsterdam)
The heritage collection of the Allard Pierson (AP) in Amsterdam includes thirteen Egyptian animal mummies, which have recently been newly investigated. The research initially approached the question of exactly which species of crocodile one of the mummies contained, and caused a bigger drive of the AP to identify and study their animal mummies. With help of the Amsterdam Medical Centre, it was possible to acquire CT scans to take an intriguing step towards digitally investigating the otherwise vulnerable mummies. In line with this project, the 4D Research Lab (4DRL) was approached to develop an application to visualise this data for museum visitors, and make it available for different material and animal specialists involved in the project. Many existing tools are intended for specialists, and are less intuitive to work with, whereas the objective of the 4DRL app was producing a visualisation to provide info to visitors of the museum, while containing interactive tools for researchers to use. Our application makes use of the capabilities of an existing web graphics library (X3DOM) as a way to render the volumetric CT data collected from the mummies, and integrate it with scans of the exterior of the mummies. During analysis and app production, the 4DRL encountered learning points and hurdles with how web graphics and CT data can be appended for the purpose of presenting cultural heritage. These are presented along with the project context as some preliminary findings.
Digital Public Archaeology with the ArchaeoTrail App
L. Loges/S. Döpper/M. Ludwig/I. Gurjanow/D-X. Ken Oehler (University of Glasgow/Goethe University Frankfurt)
Archaeological sites can offer valuable insights into the past for visitors and local audiences if they are presented well. Often, however, more information is needed than easily available to communicate this knowledge to a wider audience. This is especially true for less widely known and more remote sites that lack the funding for proper signage and knowledgeable guides. To fill this gap, and make local heritage accessible to people around the world, the ArchaeoTrail project based at Goethe University Frankfurt (www.archaeotrail.org) has developed a web portal for archaeologists and other people responsible for archaeological sites to generate self-guided, informative tours of their site using a smartphone app.
The ArchaeoTrail App is based on the successful two-fold system of MathCityMap (www.mathcitymap.eu), which entails a web portal and an actual app for mobile devices. The portal (www.archaeotrail.org) aims at archaeologists, research institutions and even volunteers working in heritage worldwide who want to create tours of archaeological sites. With just a few clicks, texts, photos, audio files and videos can be uploaded in the web portal. By doing so, a new tour is automatically created. In the smartphone app, visitors are guided by GPS along a suggested route to the stations of the tour. They receive information in the form of texts and multimedia. There is also an option for more child-friendly tours for school groups and families with fun quizzes.
Some of the first tours, or “trails”, have been generated at the site of Al-Khashbah in Central Oman, where the Frankfurt University team were able to test the application and receive feedback on the trails from local residents. The same is planned for several other sites in the Sultanate of Oman that members of the team have worked on. Other trails are planned or already published in Germany and other countries including Scotland, Namibia, Turkey, and Syria. In our talk, we will share the insights gained during this testing phase, as well as the consequences it resulted in for these and future trails.
Archaeology and digital games/simulations: a cyber-archaeological VR prototype about the Vetera I legionary camp
Matheus Morais Cruz (University of São Paulo)
In recent years, the academic community, including archaeologists, has shown great interest in the relationship between digital games (and/or simulations), research, education, and extroversion. Archaeology aims to study material culture and its physical, functional, and symbolic aspects, through specific research methodologies. Digital applications are nothing more than human products, recorded in physical or digital media and created from the fusion of four elements: programming codes, texts (computational linguistics), audio, and art (graphic design) (Copplestone 2017), being, therefore, a part of contemporary material culture loaded with meanings produced from subjective interpretations of reality by groups of people. On the other hand, games can also be defined, more broadly, as playful means of learning that go back hundreds of years, capable of providing complex approaches to the past and present, through their interpretive frameworks correlated with a contextualized use of technology. Thus, the relationship between archaeology and digital games/simulations has been the concern of several disciplines within Digital Archaeology, among them Cyber-archaeology (Forte 2010).
Based on the concepts and practices of Cyber-archaeology and the possibilities that the discipline offers for the extroversion of archaeological knowledge, this paper aims to present the ideas, methods and difficulties involved in the creation of an interactive simulation application named VETERA VR: a low-poly walking simulator, whose scenario takes place in the Vetera I legionary camp, built in the region of the modern Xanten (Germany) in 16 BC during Drusus’ campaigns against the Germanic tribes from west-central Europe, and which was completely destroyed between 69 and 70 AD during the Revolt of the Batavi.
The main objective of the application is to test the feasibility of the project for the future development of a more complete digital game, which will be one of the final products of my master’s research entitled “The amphorae of Vetera I: contacts, frontiers, and Roman military supply in the limes germanicus”, under development at the Museum of Archaeology and Ethnology (MAE-USP) at the University of Sao Paulo (Brazil), which aims to collaborate in the extroversion of research results and fundamental themes and concepts for the study of the Ancient Mediterranean to a wide audience, including students from Elementary and High Education.
As pointed out by Erik Champion (2015), these digital products offer a learning potential through tools focused on non-textual visualization, in addition to their ability to convey stories and meanings. Although games are socially received, above all, as commercial products, they also assume the role of disseminators of a certain knowledge shaped and programmed by their developers, and can be worked from an educational perspective, and, therefore, must be understood as sources, primary or secondary, to be analysed with the criticality required by scientific practice.
Memories from a digital city: Use of Archaeogaming on “Pasajeros”, a Colombian videogame
Manuel Palacio (Universidad de los Andes. Colombia/Ruhr Universität. Germany)
Since Video Games belong to the modern culture, archaeology has turned its sight over them. The most famous way to get into “Archaeology of Videogames” is archaeogaming. This approach was created by Andrew Reinhard in 2014 and it has grown on in the last years. Now, there are several academical publications on scientific Magazines and books about the topic of archaeogaming, as well as blogs, YouTube videos and interviews.
My aim in this presentation is to share and discuss the results of an archeological approach on one Colombian videogame, under the Archaeogaming frame. The game we choose is Pasajeros (2022). This game sems to be a digital memory of contemporary Bogota, since the growing of the city and the inner changes due to the Metro first line building are transforming the space of the city. In this game, you play as a taxi driver in the city of Bogota of XXI century, especially on Chapinero, a famous district in Bogota where gentrification and the hipster “new wave” occurs. As a driver you must get different people to various places of the city at night. The gameplay allows you to speak with the passengers, to know their stories and motivations, while you drive around Chapinero. According to the Archaeogaming statements, it is possible to study a game as an archeological site (Reinhard, 2018) and this thesis describes our aim.
The main objective of the presentation is to show the results of an application of Archaeogaming’s methodology on the videogame “Pasajeros” in order to identify a digital memory from Bogota, facing the current changes that take place in the city. The investigation has applied the strategies of archeogaming according to the methodology described by Reinhard (2018, p. 108-127) in five steps: 1. Digital survey; 2. archeological plan for a synthetic Bogota; 3. "excavating" a synthetic Bogota; 4. Documentation, Chronology and Location; and at least 5. synthesis and publication. We were able to apply this methodology because we have worked with a 3d representation of Chapinero, as well we did a physical comparison with the real places in the city. We could get an application of the analyses given by Reinhard about “Landscape Archaeology in Video Games” (2018, p. 95), and achieve a deeper understanding of the memories from Chapinero.
As results, we found that archaeogaming is able to give an archeological approach to digital representations of Bogota, which implies both levels at the same time: references to all-day representations of the city by its dwellers, and references to historical monuments and its signification to the citizens who live the modern changes of the city. Nevertheless, Archaeogaming’s methodology presents some difficulties, which are pointed out in the conclusion.
This research was made by researchers from UNAD in association with a group of Archeologist from Chile.
From previous on-site reconstructions to current virtual 3D reconstructions - A look at the impact of recreating the past using the example of the megalithic tomb Düwelsteene in Westphalia
Louise Tharandt (University of Cologne)
The Düwelsteene near Heiden, Westphalia, is one of the most southern megalithic tombs of the Funnel Beaker culture. In 1932 the Düwelsteene were restored and the appearance of the grave was changed. Even though the megalithic tomb was excavated under archaeological supervision and was restored based on the scientific knowledge at that time, there are not many sources of how the tomb looked like before it was restored.
In 2017 a digital twin of the megalithic tomb was created through image-based modelling. With this 3D model, the Düwelsteene were not only digitally preserved, it was also possible to reconstruct how the megalithic tomb could have looked like before the restoration in 1932. The basis of this reconstruction was a Citizen Science project, in which the citizens of Heiden, which is near the Düwelsteene, and the surrounding areas were asked to look for and send in photographs of the megalithic tomb before the restoration in 1932. With these images the digital reconstruction of the Düwelsteene before the restoration could be created. With the knowledge of the placement of some of the megaliths, it was also possible to build a virtual reconstruction of the Düwelsteene and how they could have looked like around 3000 BC, a time in which the megalithic tomb had been used, based on the pottery finds and carbon dating of animal bones.
For the master's thesis in Digital Archaeology at the University of Cologne the Düwelsteene can now be viewed as three different 3D models on a website, which was created as a visualisation of the megalithic tomb and also a supplement for the 3D models of the thesis. The "Devil's Stones" are displayed as a digital model of how the grave looks today, they are depicted as a 3D reconstruction before their restoration in 1932 and as a reconstruction model of how the megalithic tomb could have looked like around 3000 BC. All of the models have annotations and through features of the 3D viewer (3DHOP), the models can be measured or sectioned.
I will present the benefits of virtual reconstructions for cultural heritage and the advantages that combining digital archaeology with citizen science projects can have on research results as well as on the integration of the general public in archaeological projects.