Periodic Reporting for period 1 - MarginScapes (Long-term land use and water management strategies in arid margin landscapes)
Berichtszeitraum: 2018-03-19 bis 2020-03-18
MarginScapes set out to bring together large-scale, multi-temporal and multi-source geospatial analysis to re-evaluate the socio-ecological dynamics that shaped the cultural landscapes of South Asia. The project has focused on the mapping of archaeological sites in the Cholistan Desert (Pakistan) and the northern margins of the Thar desert (north-western India). Those areas were core regions for the development of the Indus Civilisation (ca. 3300-2500 BC) and have been the subject of considerable historical interest due to the presence of an extensive network of relict riverbeds that are thought to have supported several ancient sites in what is today an extreme arid ecotone.
By using a novel combination of Earth Observation data and analysis in petabyte-scale cloud computing environments, machine learning geostatistics and GIS-based topographic analysis, MarginScapes has provided new methodological tools and quantifiable open access data to: 1) understand the relationships between past hydrological systems, relict palaeosoils and the distribution of ancient sites; 2) identify historical and modern landscape transformations; and 3) foresee the mechanisms of how populations coped and adapted to climate change, water scarcity and desertification.
By using a novel combination of Earth Observation data and analysis in petabyte-scale cloud computing environments, machine learning geostatistics and GIS-based topographic analysis, MarginScapes has provided new methodological tools and quantifiable open access data to: 1) understand the relationships between past hydrological systems, relict palaeosoils and the distribution of ancient sites; 2) identify historical and modern landscape transformations; and 3) foresee the mechanisms of how populations coped and adapted to climate change, water scarcity and desertification.
The academic dissemination of the project started at the 24th conference of the European Association for South Asian Archaeology and Art (Naples, 02-06 July 2018) with a co-authored presentation that introduced the action. The ongoing action was also presented at the European Space Agency Living Planet Symposium (Milan, 13-17 September 2018), and the final results were presented at the Annual Conference Indian Society for Prehistoric and Quaternary Studies (Thiruvananthapuram, 7-9 November 2019). The project has been disseminated to academic MA and PhD scholars in the International Workshop on GIS, Remote Sensing and Geoarchaeology (Department of Archaeology, University of Kerala, Thiruvananthapuram, 1-7 December 2019), and at the Host Institution, the McDonald Institute for Archaeological Research, University of Cambridge, the research was presented at the IberoArchUK 2019 Conference, the PostDoc Forum and the Cambridge Science Festival.
A core part of the project has been the training and application of computational methodological workflows towards: a) the synergistic integration of Earth Observation sensors (radar and multispectral); and b) the automated detection of features of archaeological, historical, and geomorphological interest. Under the supervision of Drs Petrie and Orengo, the PI has integrated data from the European Space Agency (i.e. Sentinel 1 and Sentinel 2 satellite image collections) in new processing workflows that can be used for the detection of archaeological sites in arid landscapes. This constitutes an important new development, as recent research has emphasized the need for the integration of distinct sensors, in what is known as a “virtual constellation”. MarginScapes has led to the detection and accurate location of c. 500 archaeological mounds in the Cholistan Desert, documenting many more archaeological mounds than previously recorded, extending south and east into the desert, which has major implications for understanding the archaeological significance of the region (paper submitted to PNAS). This data has been complemented with high-resolution WorldView satellite imagery granted by DigitalGlobe Foundation. Dr Conesa has also participated in specific Earth Engine training events, such as the Google Geo for Good User Summit (Sunnyvale, 1-5 October 2018) and the Google Earth Engine User Summit (Dublin, 12-14 June 2018, see Tworains Blog for a review: tworains.wordpress.com).
The project has also integrated a set of historical topographical maps, including a) collections from the Survey of India maps (1880s-1947); and b) the 1:250K US Army Map Services published in 1955. These historical map series have proven useful in the evaluation of landscape change and the detection of archaeological features. The maps have been acquired through visits at the Map Room, Cambridge University Library, and the British Library Map collections. This work has contributed to the joint paper led by Dr Petrie and the TwoRains team entitled “Mapping Archaeology While Mapping an Empire: Using Historical Maps to Reconstruct Ancient Settlement Landscapes in Modern India and Pakistan” (Geosciences, 2019).
MarginScapes have also re-evaluated the historical and heritage significance of c. 40 historical fortifications between Pakistan and India. Some of them appear in the UNESCO Tentative List, but have received little attention. Using those forts as a baseline, the research has reconstructed the long-term socio-ecological dynamics that shaped the vibrant merchant and nomadic lifestyles of the Cholistan Desert for centuries. Using a combination of satellite imagery, GIS-based analysis and the use of historical sources, the work has delineated historical trade routes that show a close correspondence with the presence of seasonal water bodies (manuscript in preparation).
Last but not least, MarginScapes has culminated with the organisation of an international workshop at the Host Institution. The workshop, entitled Computational approaches to archaeological site detection and monitoring, brought together a group of recognised experts in the application of remote sensing applications to archaeological research, with a specific focus on site location methods, to share procedures and experience that can serve to direct future applications under a common ground. More information about the workshop can be found in the event website (archaeositedetection.wordpress.com) and the papers that were presented will be published in a co-edited volume.
A core part of the project has been the training and application of computational methodological workflows towards: a) the synergistic integration of Earth Observation sensors (radar and multispectral); and b) the automated detection of features of archaeological, historical, and geomorphological interest. Under the supervision of Drs Petrie and Orengo, the PI has integrated data from the European Space Agency (i.e. Sentinel 1 and Sentinel 2 satellite image collections) in new processing workflows that can be used for the detection of archaeological sites in arid landscapes. This constitutes an important new development, as recent research has emphasized the need for the integration of distinct sensors, in what is known as a “virtual constellation”. MarginScapes has led to the detection and accurate location of c. 500 archaeological mounds in the Cholistan Desert, documenting many more archaeological mounds than previously recorded, extending south and east into the desert, which has major implications for understanding the archaeological significance of the region (paper submitted to PNAS). This data has been complemented with high-resolution WorldView satellite imagery granted by DigitalGlobe Foundation. Dr Conesa has also participated in specific Earth Engine training events, such as the Google Geo for Good User Summit (Sunnyvale, 1-5 October 2018) and the Google Earth Engine User Summit (Dublin, 12-14 June 2018, see Tworains Blog for a review: tworains.wordpress.com).
The project has also integrated a set of historical topographical maps, including a) collections from the Survey of India maps (1880s-1947); and b) the 1:250K US Army Map Services published in 1955. These historical map series have proven useful in the evaluation of landscape change and the detection of archaeological features. The maps have been acquired through visits at the Map Room, Cambridge University Library, and the British Library Map collections. This work has contributed to the joint paper led by Dr Petrie and the TwoRains team entitled “Mapping Archaeology While Mapping an Empire: Using Historical Maps to Reconstruct Ancient Settlement Landscapes in Modern India and Pakistan” (Geosciences, 2019).
MarginScapes have also re-evaluated the historical and heritage significance of c. 40 historical fortifications between Pakistan and India. Some of them appear in the UNESCO Tentative List, but have received little attention. Using those forts as a baseline, the research has reconstructed the long-term socio-ecological dynamics that shaped the vibrant merchant and nomadic lifestyles of the Cholistan Desert for centuries. Using a combination of satellite imagery, GIS-based analysis and the use of historical sources, the work has delineated historical trade routes that show a close correspondence with the presence of seasonal water bodies (manuscript in preparation).
Last but not least, MarginScapes has culminated with the organisation of an international workshop at the Host Institution. The workshop, entitled Computational approaches to archaeological site detection and monitoring, brought together a group of recognised experts in the application of remote sensing applications to archaeological research, with a specific focus on site location methods, to share procedures and experience that can serve to direct future applications under a common ground. More information about the workshop can be found in the event website (archaeositedetection.wordpress.com) and the papers that were presented will be published in a co-edited volume.
The methods and datasets generated by the project represent a step forward for the archaeology of South Asia, in particular for the understanding of the Indus Civilisation, and it contributes to the field of remote sensing applications in archaeological research. This is particularly relevant nowadays when machine-learning applications (and, more broadly, artificial intelligence and Big Data) are now at the forefront of archaeological applications and site detection workflows.
The review and improvement of old archaeological legacy data, as well as the reassessment of the archaeological and cultural heritage in the Cholistan Desert, constitutes a new corpus of information that can be used for stakeholders and cultural heritage organisations, thus contributing to future research and actions towards the preservation of neglected heritage. To this end, all data, code and algorithms will be made public an open-access. For example, a planned publication that presents the detailed findings of the research in Cholistan will be submitted to the Journal of Open Archaeological Data.
The review and improvement of old archaeological legacy data, as well as the reassessment of the archaeological and cultural heritage in the Cholistan Desert, constitutes a new corpus of information that can be used for stakeholders and cultural heritage organisations, thus contributing to future research and actions towards the preservation of neglected heritage. To this end, all data, code and algorithms will be made public an open-access. For example, a planned publication that presents the detailed findings of the research in Cholistan will be submitted to the Journal of Open Archaeological Data.