One more year, multiple GIAP (ICAC) researchers will be presenting papers and co-organising a session at the CAA2023 (Annual Conference of Computer Applications and Quantitative Methods in Archaeology) that will be held from April 3rd to 6th 2023 in Amsterdam (Netherlands).
Here’s a list of contributions (click to navigate to the abstracts below)
- Paper 115: ‘Near East irrigation studies beyond SRTM and CORONA. Preliminary results of the UnderTheSands project‘
- Nazarij Buławka et al.
- Paper 136: ‘Convolutional multi-factor probabilistic time-aware corridors: a new approach to the analysis of past long-distance mobility‘
- Paper 137: ‘From shape to grow conditions: a workflow combining micro-3D scanning, geometric morphometrics and machine learning for the analysis of past agricultural strategies‘
- Hèctor A. Orengo et al.
- Hèctor A. Orengo et al.
- Paper 220: ‘Loss of heritage in riverine monsoonal plains: integration of geospatial legacy data and multi-source remote sensing in Jammu, India‘
- Navjot Kour et al.
- Navjot Kour et al.
- Paper 264: ‘Machine Learning approaches for a multi-scale and multi-source detection and characterization of archaeological sites: the case of the funerary tumuli at Abdera (Thrace, Greece, 6th – 2nd C. BCE, aprox.)‘
- Arnau Garcia-Molsosa et al.
- Arnau Garcia-Molsosa et al.
- Paper 333: ‘Current approaches to computational archaeology at GIAP (ICAC)‘
- Hèctor A. Orengo et al.
- Hèctor A. Orengo et al.
- Paper 286: ‘Mapping Indus archaeology and multi-temporal land cover trends in semi-arid regions‘
- Francesc C. Conesa et al.
- Paper 359: ‘Mapping Archaeological Heritage in South Asia: a digital workflow in development‘
- Rebecca C Roberts (University of Cambridge) et al.
- Rebecca C Roberts (University of Cambridge) et al.
- Session 06: Stay connected: Developing Mobile GIS for team-based collaboration in archaeological research
-Organisers: Julia M. CHYLA, Adéla SOBOTKOVA, Nazarij BUŁAWKA (ICAC), Giuseppe Prospero CIRIGLIANO
- Poster 363: The preliminary results of application of remote sensing and machine learning methods in the studies of flint mining landscape in Egypt
– Nazarij Buławka
Convolutional multi-factor probabilistic time-aware corridors: a new approach to the analysis of past long-distance mobility
Type: Paper, paper number: 136
Authors: Hector A. Orengo (ICREA Research Professor at the Landscape Archaeology Research Group, Catalan Institute of Classical Archaeology)*; Toby C Wilkinson (Catalan Institute of Classical Archaeology) and Cameron A. Petrie (McDonald Institute for Archaeological Research of the University of Cambridge)
Mobility has been essential for the development of communications and relationships between different human groups. Transcontinental routes, formed by the continuous adjoining of intra-cultural roads, played an essential role in the development of a common knowledge of the world as early as the Bronze Age with clear evidence of cultural interchange between lands so distant as Crete in the Mediterranean and India. These are shaped by interlinked settlements and extend over territories surpassing the regional scale. When approached at a large scale routes join forming transport networks forming a continuous web of movement possibilities, which expands in space and time.
Despite the importance of movement to comprehend the origins and development of human relationships little is understood of how people moved in the past, how environmental conditions affected movement and how the domestication of animals and the evolving technologies of transport extended human connectivity to different ecological zones. This is particularly pressing when considering long-distance and transcontinental routes, where factors affecting movement can greatly vary along the route and the mobility network tend to reflect seasonal variations and socio-economic factors.
In this paper we address these problems by developing a network of high resolution transcontinental, multi-factor probabilistic temporal corridors using high performance computing that can be statistically related to multitemporal settlement data and queried using common statistical approaches.
Current approaches to computational archaeology at GIAP (ICAC)
Type: Paper, paper number: 333
Authors: Hector A. Orengo (ICREA Research Professor at the Landscape Archaeology Research Group, Catalan Institute of Classical Archaeology)*
This paper will present current development strategies, theoretical approaches, and practical workflows currently in use at the Landscape Archaeology Research Group, Catalan Institute of Classical Archaeology.
Development strategies will be explained as an evolving compromise between group development and the career development of the researchers within the group. These will showcase the strategies followed to keep the members’ work integrated through collaborative work in complementary projects. These projects cover specific research areas within the main computational research lines promoted by the group leaders.
Theoretical approaches adopted by the group will be explained in relation to the group development strategies, the research lines promoted and the practical approaches adopted for each specific project.
Lastly, practical workflows will be explained in relation to specific research projects. The different toolboxes and workflows available to members of the group are combined, expanded and complemented in order to achieve the specific objectives of each project. Several specific workflows characteristic of our research group in areas such as landscape archaeology, archaeological survey, site detection and monitoring, machine learning and 3D object analysis will be showcased in certain detail.
The presentation will make extensive use of maps, flow diagrams and other charts as recommended by the session organisers to illustrate the current approaches to computational archaeology by the GIAP.
This paper will discuss the theoretical assumptions, practical approaches and the research environment that underpin and frame current approaches to computational archaeology at GIAP. As such it will be both subjective and explicit in geolocating current computational research within the frame of the current academic market, funding landscape, and emerging digital trends in archaeology and the humanities.
From shape to grow conditions: a workflow combining micro-3D scanning, geometric morphometrics and machine learning for the analysis of past agricultural strategies.
Type: Paper, paper number: 137
Authors: Hector A. Orengo (ICREA Research Professor at the Landscape Archaeology Research Group, Catalan Institute of Classical Archaeology)*; Alexandra Livarda (Landscape Archaeology Research Group, Catalan Institute of Classical Archaeology); Alexandra Kriti (Landscape Archaeology Research Group, Catalan Institute of Classical Archaeology); Ioannis Mylonas (Institute of Plant Breeding & Genetic Resources); Elissabet Ninou (Department of Agriculture, International Hellenic University)
The shape of objects encloses information about their function. Geometric morphometrics (GM) in combination with statistical approaches have been employed to obtain quantifiable data about objects’ shapes and use these data to define typologies and functionalities. Equally, animal bones (including human) and archaeobotanical remains have been analysed using similar approaches to obtain information, compare specimens and classify items.
During the last years the increased availability of 3D scanners and more efficient photogrammetry algorithms and software have facilitated the generation of 3D models of archaeological items. However, despite the multiplication of 3D models, little has been done to develop new workflows that can take advantage of the volumetric nature of these models and simple GM measures continue to be employed to define complex objects. Similarly, the questions under investigation have not evolved to take advantage of the detailed information that 3D models offer. This is even more evident in the case of archaeobotanical data, the small size (typically sub-centimetric) of which restricts their accurate 3D scanning.
In this paper we will present current ongoing work to develop a workflow based on a combination of micro-3D scanning, 3D GM and machine learning using archaeobotanical material. The objective is to go beyond the simple identification of seeds, to try identify the growing conditions of archaeological grains, which likely include the agricultural regimes employed by past societies.
Loss of heritage in riverine monsoonal plains: integration of geospatial legacy data and multi-source remote sensing in Jammu, India
Type: Paper, paper number: 220
Authors: Navjot Kour (Institut Català d’Arqueologia Clàssica)*; Francesc Conesa (Catalan Institute of Classical Archaeology); Arnau Garcia-Molsosa (Landscape Archaeology Research Group, Catalan Institute of Classical Archaeology); Hector A. Orengo (ICREA Research Professor at the Landscape Archaeology Research Group, Catalan Institute of Classical Archaeology)
Floodplains are one of the most visible and widespread landforms on Earth’s surface. Over time they develop and change, and so they continuously reshape their river systems and have a major impact on short- to longer-term human-environment interactions. Most contemporary riverine landscapes, if not all, have been altered by anthropic influence throughout the Holocene. In recent years, however, the intensification of human activities has increased the rate at which natural processes shape the landscape and have severely aggravated the preservation and visibility of many archaeological and cultural heritage sites. In such contexts, remote sensing procedures become an essential tool to remotely investigate large-scale archaeological patterns, land use trends and landscape change at multiple scales of observations.
In recent years, we have witnessed a growing number of geospatial and remote-based applications in South Asian archaeological landscapes (e.g., Conesa et al. 2015, Orengo and Petrie 2017, Orengo et al. 2020, Rajani 2020). This thriving trend is placing South Asian landscapes and in particular the areas within the influence of the Indus Civilisation and the greater Indus Valley (aprox. 3500 BC – 1900 BC), as optimal land laboratories for the development of new methods and workflows that can be exported to and compared with other similar geographical areas with a long-standing tradition of remote-based investigations, such as the Near East (e.g., Rayne et al. 2020).
Here we present the first steps of RIVERINE, a new collaborative project that revolves around the interactions between the archaeology and the hydrological resources of the Jammu alluvial plains in north-western India. We aim to detect and map the location of ancient mounds and re-evaluate their cultural and chronological contexts while also exploring the historical and present-day land use trends and anthropogenic disturbances that put the Cultural Heritage of Jammu at risk.
Machine Learning approaches for a multi-scale and multi-source detection and characterization of archaeological sites: the case of the funerary tumuli at Abdera (Thrace, Greece, 6th – 2nd C. BCE, aprox.)
Type: Paper, paper number: 264
Authors: Arnau Garcia-Molsosa (Landscape Archaeology Research Group, Catalan Institute of Classical Archaeology)*; Hector A. Orengo (ICREA Research Professor at the Landscape Archaeology Research Group, Catalan Institute of Classical Archaeology); Mercouris Georgiadis (Landscape Archaeology Research Group, Catalan Institute of Classical Archaeology); Nikolas Dimakis (National & Kapodistrian University of Athens); Paraskevi Motsiou (National & Kapodistrian University of Athens); Alfredo Mayoral (Landscape Archaeology Research Group, Catalan Institute of Classical Archaeology); Eurydice Kefalidou ( National & Kapodistrian University of Athens); Konstantina Kallintzi (Eforate for Antiquities of Xanthi, Greek Ministry of Culture)
Machine and Deep Learning methods are being successfully applied for the detection of topographic anomalies of anthropogenic origin. After the publication of initial proofs of concept, applications addressing different study cases are quickly growing during the last years. First applications tend to focus on the analysis of a single source (usually a EO product of a particular LIDAR dataset), but the potential of multi-source analysis is being explored in recent works. In most of the examples, ML and DL have been implemented as a “brute force” instrument, particularly suited to analyse very extensive areas which would be impossible to cover by traditional manual means (both on field and on computer image analysis), but providing little value for local-scale analysis. As a consequence, the target of these analyses are dependent of the availability and resolution of the sources employed. That has directed most of the analysis towards metric-resolution LIDAR national or regional datasets (where those are available), while global satellite-based sensors have been employed to locate large mounded-sites in areas such the Middle-East or the Indus Valley. The potential of higher resolution imagery for targeting all sorts of archaeological features (from potsherds to little mounds or crop-marks), is being tested with success using UAS-mounted sensors in limited areas.
Results obtained through automatic site-detection address also a unique characteristic, most commonly the topographical anomaly. Multi-source analyses have incorporated other factors, but employed basically to distinguish sites of other similar signatures. Although the indisputable value of the location of hundreds or thousands of potential new sites, it represents only a very first step on the research, with very limited historic-archaeological interpretative value.
In this paper we will address the potential of integrating multi-source and multi-scale approaches to provide information beyond the location of potential archaeological sites including some elements of its historic-archaeological character. To achieve this objective, we will use the study case of the funerary tumuli spread on the hinterland of the Greek colony of Abdera (Thrace), founded in the 7th c. BCE and occupied until the Byzantine period. The area is nowadays under intensive cultivation, which has involved the levelling of the tumuli, resulting in few of them still standing, and most of them recognisable only by as small elevations in the fields associated to remains of sarcophagi and stones. Funerary tumuli in this area had been the focus of previous works, including some recent work lead by the authors of the paper, and its distribution is well-known. That made it a particular adequate case to test our workflow.
Mapping Archaeological Heritage in South Asia: a digital workflow in development
Type: Paper, paper number: 359
Rebecca C Roberts, University of Cambridge; Junaid Abdul Jabbar, University of Cambridge; Iban Berganzo-Besga, Institut Català d’Arqueologia Clàssica, ICAC; Rosie Campbell, University of Cambridge; Moazzam Durrani, Islamia University Bahawalpur; Arnau Garcia-Molsosa, Landscape Archaeology Research Group, Catalan Institute of Classical Archaeology; Petrus Gerrits, University of Cambridge; Jonas Gregorio, Universitat Pompeu Fabra; Abhayan G.S., University of Kerala; Muhammed Hameed, University of the Punjab; Afifa Khan, University of Cambridge; Marco Madella, Universitat Pompeu Fabra; Hèctor A. Orengo, Landscape Archaeology Research Group, Catalan Institute of Classical Archaeology; V.N. Prabhakar, IIT Gandhinagar; Abdul Samad, Directorate Of Archaeology & Museums, Khyber Pakhtunkhwa; Rajesh Sasidharan, University of Kerala; Ravindra Nath Singh, Banaras Hindu University; Vikas K. Singh, Banaras Hindu University; Maria Suarez Moreno, University of Cambridge; Kuili Suganya, University of Cambridge; Jack Tomaney, University of Cambridge; Azadeh Vafadari, University of Cambridge; Muhammed Waqar Mushtaq, Islamia University Bahawalpur; Cameron Petrie, University of Cambridge
The Mapping Archaeological Heritage in South Asia (MAHSA) project is developing heritage management tools and expertise that support systematic documentation of archaeological heritage in Pakistan and northwest India. We are developing tools to document the full range of archaeological heritage, which span in date from the earliest villages, through several phases of urbanism, the rise and fall of numerous historical states and empires, and up to the colonial and modern periods. Today, many areas are densely occupied and undergoing rapid development. Many sites are at risk, typically from factors including erosion, large-scale development, looting, and the expansion of extensive irrigation agriculture and the concomitant levelling of large tracts of land. Site destruction has been observed in the field and is ongoing, and at present the level and rate of site loss is not being monitored in many areas. There is thus a clear need for archaeological sites to be documented in a system that also enables monitoring and management (Gupta et al. 2017) This paper explores the work and challenges of the Mapping Archaeological Heritage in South Asia (MAHSA) project in using different technologies and methods to record, digitise, re-interpret, link, and re-use non-digital and born-digital archaeological data into a structured and standardised digital format using a common and unique controlled vocabulary, so that it may become findable, accessible, interoperable, and reusable (FAIR) data.
Mapping Indus archaeology and multi-temporal land cover trends in semi-arid regions
Type: Paper, paper number: 286
Authors: Francesc Conesa (Catalan Institute of Classical Archaeology)*; Hector A. Orengo (ICREA Research Professor at the Landscape Archaeology Research Group, Catalan Institute of Classical Archaeology); Arnau Garcia-Molsosa (Landscape Archaeology Research Group, Catalan Institute of Classical Archaeology); Abhayan GS (University of Kerala); Rajesh SV (University of Kerala)
This paper will present ongoing remote-based research in a selection of South Asian Indus contexts ranging from the Kachchh region in north-western India to the arid landscapes of the Cholistan Desert in Pakistan. These regions were areas of significant influence for the development of the Indus Civilisation and saw the flourishment of villages, cities and fortified citadels which coexisted with agro-pastoral communities dedicated to the exploitation of local resources. In Kachchh, in particular, newly excavated sites by the University of Kerala, such as the cemetery of Juna Khatiya, attest to the first communities showing Indus traits (c. 3300-2600 BC). Later in historical and medieval times, a network of forts and caravan routes also indicate the long-term relevance of these areas in large-scale mobility and trade connections within the greater Indus region and beyond, towards Central Asia and the Gulf.
In recent years, remote sensing applications for archaeology have seen an unprecedented advance in the application of computational approaches for the remote analysis of archaeological landscapes and features. This progress is mostly related to improvements in availability (i.e. open access), diversity and quality of new satellite missions and the increasing implementation of cloud data catalogues and online computing and visualisation platforms. Within this context, we aim to re-evaluate the archaeological cultural landscapes of South Asian drylands, by 1) expanding the current archaeological data gazetteers with the detection and mapping of mounds and sites, and 2) evaluating the potential risks (from geohazards to anthropic impact) affecting its preservation, and 3) understand its location and position in relation to the potential land use available resources and landscape connectivity through time.
Near East irrigation studies beyond SRTM and CORONA. Preliminary results of the UnderTheSands project
Type: Paper, paper number: 115
Authors: Nazarij Bulawka (Landscape Archaeology Research Group, Catalan Institute of Classical Archaeology)*; Hector A. Orengo (Landscape Archaeology Research Group, Catalan Institute of Classical Archaeology)
Wittfogel’s hydraulic hypothesis (1957) motivated much of the early archaeological discourse in the Near East to seek a connection between large-scale irrigation and the centralized power in early states; provoking researchers to challenge his views. The current research shows a more complicated picture2. Various agencies controlled the irrigation works, but water management was indeed the backbone of ancient Near Eastern society.
Advances in irrigation research have used multiple methods, from historical analyses to the use of GIS. In terms of remote sensing, LANDSAT (Adams 1981), CORONA and SRTM satellite imagery paved the way for large-scale landscape studies in the Near East (Ur 2002; Philip et al. 2002; Wilkinson 2003). Unfortunately, because of the political situation, most of the research on irrigation has almost stopped in recent years or moved elsewhere. The importance of irrigation studies has not diminished but its magnitude has rather increased due to current environmental and political issues.
The paper will present the UnderTheSands project, aiming to locate and reconstruct the irrigation network of the areas under study in the Near East and explore their chronological dynamics. Four areas have been selected for the project. These are located in Iraq, Iran and Turkmenistan. The chosen regions in UnderTheSands have diversified characteristics, which will allow to development of a novel workflow for the large-scale analysis of irrigation networks that later can be applied to any other areas presenting similar features. By studying the irrigation networks in these selected areas, we will understand the long-term socio-economic and historical circumstances that produced and maintained these irrigation networks.
The proposed methodology will include remote sensing, terrain analysis, hybrid Machine / Deep Learning methods, archaeomorphology, spatial correlation indices, and historical studies) and various spatial data sources (including multispectral imaging, synthetic aperture radar, and TanDEM-X). The application of Google Earth Engine (GEE) will create a possibility to work with an enormous amount of data incorporating the workflow developed by UnderTheSands. The project will be based at the Landscape Archaeology Research Group (GIAP) at the Catalan Institute of Classical Archaeology (ICAC) in Tarragona.
The paper will present the research objectives, methods and initial results of the “UnderTheSands: Ancient irrigation detection and analysis using advanced remote sensing methods” project (HORIZON-MSCA-2021-PF-01-101062705), funded by The European Union.
- Adams, Robert McCormick. 1981. Heartland of Cities. Surveys of Ancient Settlement and Land Use on the Central Floodplain of the Euphrates. Chicago / London: University of Chicago Press.
- Philip, G, D Donoghue, A Beck, and N Galiatsatos. 2002. “CORONA Satellite Photography: An Archaeological Application from the Middle East.” Antiquity 76 (291): 109–18.
- Ur, Jason A. 2002. “Surface Collection and Offsite Studies at Tell Hamoukar, 1999.” Iraq 64: 15–43. https://doi.org/10.2307/4200517.
- Wilkinson, Tony James. 2003. Archaeological Landscapes of the Near East. Tucson: University of Arizona Press.
- Wittfogel, Karl August. 1957. Oriental Despotism; a Comparative Study of Total Power. New Haven: Yale University Press.
The preliminary results of application of remote sensing and machine learning methods in the studies of flint mining landscape in Egypt
Authors: Nazarij Bulawka (Landscape Archaeology Research Group, Catalan Institute of Classical Archaeology)
The archaeological studies of flint mining are a vast topic in prehistoric research worldwide. A variety of methods have been used. Many of them focus on the studies of the unique relief of terrain in areas where flint was extracted.
Egypt is ideally situated to examine both natural and human-made terrain changes in the landscape. It has extensive desert regions devoid of or merely sparsely vegetated, and many such areas are intact. Taking that, the search for flint extraction points can be particularly promising with a computational or remote sensing approach.
Flint raw material deposits were found in Egypt between Esna in the south and Cairo in the north, along the wadis near the Nile Valley. Archaeological investigations have been carried out at several flint mines in Egypt. Among the best known are Nazlet Khater, Nazlet Safaha, Wadi el-Sheikh or the Eastern Desert mines (Vermeersch 2002; Briois and Midant-Reynes 2014; Köhler, Hart, and Klauzner 2017). From the Palaeolithic through the New Kingdom, their use has been dated differently at each location. These archaeological sites come in great numbers, many of which have not yet been located. The studies conducted so far did not allow an understanding of the extent of the mining activities.
The paper will present the result of the application of remote sensing, Machine and Deep learning methods in the studies of flint mining in Egypt. Two test areas have been selected for the studies, representing the best-preserved mining landscape.
The data used in the project are high-resolution WorldView, Google Earth Pro, Sentinel 2 multispectral imagery and Copernicus Digital Elevation Model (DEM) (GLO-30), which were selected because they represent the best resolution among freely available data.
The paper will present the preliminary results of the project and discuss the potential and issues in the study of flint mining on the selected examples in Egypt. The plans for future projects will also be discussed.
- Briois, François, and Béatrix Midant-Reynes. 2014. “Sur Les Traces de Georg August Schweinfurth: Les Sites d’exploitation Du Silex d’époque Pharaonique Dans Le Massif Du Galâlâ Nord (Désert Oriental).” Bulletin de l’Institut Français d’Archéologie Orientale. PP -.
- Köhler, E Christiana, Elizabeth Hart, and Michael Klauzner. 2017. “Wadi El-Sheikh: A New Archaeological Investigation of Ancient Egyptian Chert Mines.” PLoS One. PP -. https://doi.org/10.1371/journal.pone.0170840.
- Parcak, Sarah, David Gathings, Chase Childs, Greg Mumford, and Eric Cline. 2016. “Satellite Evidence of Archaeological Site Looting in Egypt: 2002–2013.” Antiquity 90 (349): 188–205. https://doi.org/10.15184/aqy.2016.1.
- Vermeersch, Pierre M. 2002. “Palaeolithic Quarrying Sites in Upper and Middle Egypt.” Egyptian Prehistory Monographs. Leuven: Leuven University Press PP – Leuven.
S06. Stay connected: Developing Mobile GIS for team-based collaboration in archaeological research
Session type: other
Organisers: Julia M. CHYLA, Adéla SOBOTKOVA, Nazarij BULAWKA (ICAC), Giuseppe Prospero CIRIGLIANO
Mobile GIS Special Interest Group has in its previous CAA conference editions (2017, 2018, 2019, 2021) drawn attention to the importance of mobile GIS in archaeological and other field research, specifically its impact and contribution to fieldwork methodology and data collection (Buławka and Chyla, 2020; Sobotkova et al., 2021).
In this edition, we survey how the use of mobile GIS in archaeology has progressed in recent years. In the early phase of COVID pandemic, Scerri et al. argued that sciences working in the field, including archaeology, had to change their ways (2021). The review of the recently published literature partly confirms it. As most international expeditions were canceled, scholars working in the field had to stop their projects. Many projects abandoned field work in favor of office work, for example remote sensing and data analysis. Others, in the late phase of the pandemic, found their way to continue working by novel methods of collaboration (Geser 2021; Magnani et al. 2021; Matte and Ulm 2021).
COVID demonstrated the benefit of producing FAIR digital data in the field. Robust toolkit ensured that the collected data were born-digital, complete and consistent upon departure from fieldwork. Having all data shared and accessible by all team-members afterwards meant that work could continue remotely, which was a source of relief during the lockdown (Sobotkova et al., 2021).
Additionally, we would like to explore the current differences in collaborative solutions between open-source and commercial software. Do the different OS and commercial software entail a different organization of archaeological fieldwork? What aspects and reasons lead a research team to choose an OS versus a commercial software? What is the range of funding models used to develop or deploy different mobile data capture applications? How should archaeologists prepare for OS updates, changes in hardware compatibility or application funding models so as to retain the ability to use the same workflow in the future?
Another aspect that we would like to address at CAA 2023 is the capability to produce standardized results that follow good practice using mobile GIS nowadays. Some archaeologists report increased fieldwork efficiency thanks to the use of mobile devices (Austin 2014; Ames et al. 2020) while others focus on the downsides of the digital medium, such as deskilling, (Caraher 2016; Gordon et al. 2016), or stress the need to manage workplace change and fine-tune daily operation under the new circumstances (Vanvalkenburgh 2018). What makes the difference? Have the mobile devices transformed the entire lifecycle of archaeological research from team-based field data capture to analysis, sharing, and publishing, or affected only the day-to-day working processes in the field? And more specifically: is mobile GIS essential for digital fieldwork? If so, what are the must-have features of mobile GIS and how do you prioritize them?