Featured

Five years on: Reflections on mobile GIS survey at Hatnub Quarry P in 2017

Map showing quarries and trails in the desert east of Amarna, including the study area and Hatnub quarries P, R and T.
Location of Hatnub and the study area in the desert east of Amarna. ((Underlying Landsat 8 data from the United States Geological Survey)

When this post goes out it will have been five years since I last worked in Egypt and almost one year since I published the results of that last season at Hatnub in the Journal of Egyptian Archaeology. In the course of writing that paper, I had the opportunity to review the results of the work. The accepted version of the paper can be found here for those who would like to read the full version.

In 2016 I remotely surveyed 1 km2 of the industrial landscape along the ancient quarry road between the Hatnub quarries and the low desert. I used high (0.4 m) resolution Worldview-3 satellite imagery projected in a geographic information system to record all the archaeological features within the study area. The results of this remote survey and comparisons with the archaeological survey undertaken by Shaw in the late 1980s were subsequently published in a peer-reviewed conference proceedings (available here). In September 2017 I returned to the site as part of the  Institut Français d’Archéologie Orientale and the University of Liverpool Hatnub Project, to undertake ground-check of the remote survey and determine its accuracy and suitability for future research projects.

Photo of the author during the survey, with tablet in hand and the Hatnub desert in the background.
Undertaking mobile-GIS survey at Hatnub in 2017. (Author photograph)

Mobile-GIS

Mobile GIS applications allow you to take your GIS project, or selected components thereof, out into the field on a tablet or similar handheld device. A mobile-GIS survey was perfect for my 2017 ground check. I uploaded the Worldview-3 satellite imagery and the remote survey data onto ArcGIS Collector (now deprecated) on my tablet. Using the tablet’s GPS software I was able to locate myself within the satellite imagery, identify survey squares, correct remote-survey data, and add additional features missed during the remote survey. I subsequently determined the percentages of false positives (i.e. features identified during the remote-survey as archaeological but found to be otherwise during the mobile-GIS survey) and false negatives (i.e. real archaeological features that were missed during the remote-survey) by comparing the original remote survey data to the mobile-GIS survey data. Careful examination of the data also revealed some likely causes for the false positives and negatives and suggested possible reasons why the results differed from the survey undertaken by Shaw in the late 1980s.

Results

Over 7 days I covered 81 grid squares, or 810,000 m2, 49 of which had been previously remotely surveyed. I recorded 573 features, including 62 linears (roads, trails, paths etc.). Habitations (huts, shelters and windbreaks), remain the largest group of features, but there are also ‘blank areas’, cairns, roads, tracks and paths, work areas, shrines, and quarries. During the course of the mobile-GIS survey, the ‘blank areas’ were reinterpreted as small quarry pits, which had filled up with wind-blown sand after the stone was removed. The mobile-GIS survey also revealed several additional types of features, which had not been recorded on the remote survey often because they were too small to be visible or appeared as natural features.

Map showing part of the study area around Quarry P, divided into survey squares, with archaeological features recorded as symbolised point data.
The Mobile-GIS survey 2017; areas surveyed, and archaeological features recorded around Quarry P (Worldview-3 imagery © 2016 DigitalGlobe Inc. supplied by European Space Imaging. Reproduced with permission).

Overall the remote survey had located 69% of the archaeological features recorded during the mobile-GIS survey. Of the features recorded by the remote survey, there were 5% false positives and 35% false negatives (these figures do not add up to 100% because the false positives slightly offset the false negatives in the overall figure). The false positives and negatives were largely the result of modern activity, ancient quarrying, and natural phenomena that obscured the desert surface. Only six were in clear areas of the desert surface, and only two of those were large enough to be seen in the satellite imagery. The false positives were particularly associated with modern activity, while ancient quarrying and natural phenomena also contributed to the false negatives.

A part of the desert where the stony surface is interrupted by a blank area of sand, which has been dug into, revealing the bedrock at the bottom.
Modern hole in a ‘blank area’, showing its irregular fill and traces of the underlying bedrock at the bottom. (Author photograph).

The rapidity with which the mobile-GIS survey recorded 573 features reveals the potential of these new technologies. Mobile-GIS survey was more thorough than the remote survey and much faster than the Shaw survey of the late 1980s.

Conclusion

The speed of the mobile-GIS survey derives from its combination of systematic remote survey with on-site updates that enhance its accuracy. It would not be possible without the twin technological developments of high-resolution satellite imagery and mobile GIS. Having undertaken traditional archaeological survey with a total station and mobile-GIS survey, the latter is definitely far faster and more efficient, with much less bulky equipment. Mobile-GIS survey also reduces the time to download, import, and process archaeological data from the total station to the computer and survey software. Unfortunately, we will need some further improvements in GPS accuracy, interoperability, and recording systems before we can entirely ‘ditch that total station‘. Excavation recording is currently too detailed for the combined precision of tablet and phone GPS and satellite imagery resolution and will continue to require additional equipment for some time to come. It may be possible to use external GPS devices to improve the accuracy of mobile-GIS, but this will be complicated by questions of interoperability between GPS, tablet and software, and whether maximum possible GPS accuracies are achievable in a specific country.

Part of the Hatnub desert. In the distance is a low hill crowned by a cairn, below that is a darker stony surface with several larger stones set upright, in the foreground is a shallow valley containing two habitation structures with various rooms.
Typical archaeological features at Hatnub. A cairn on a hill (background), a vertical standing stone set upright (left of the image), and two multi-room huts in the centre foreground. (Author photograph)

Acknowledgements

Maps and images throughout this blog post were created using ArcGIS® software by Esri. ArcGIS® and ArcMap™ are the intellectual property of Esri and are used herein under license. Copyright © Esri. All rights reserved. For more information about Esri® software, please visit http://www.esri.com.

The Worldview-3 satellite imagery is a 4-band, pan-sharpened 0.4 m resolution satellite image created from blue, green, red, and near infra-red1 multi-spectral bands pan-sharpened with the panchromatic band of the DigitalGlobe Worldview-3 satellite recorded on 9 June 2016.

The author is grateful to the Egyptian Ministry of Antiquities for permitting and facilitating the work at Hatnub. I particularly wish to thank Mr. Mohammed Khallaf, director of the Middle Egypt at the office for antiquities in Miniah, represented by Mahmoud Salah, director of Miniah at the office for antiquities in Miniah, Mr. Hamada M. Abdel Moeen Kellawy, and Mr. Mohammed Khalil, our MSA inspector for his support and his work during the fieldwork at Hatnub. The author is also grateful to Roland Enmarch for permission to use his photographs in this paper and to Yannis Gourdon and the Institut Français d’Archéologie Orientale, our collaborators on the Hatnub Project.

The purchase of the Worldview-3 satellite imagery and the 2017 ground-truthing fieldwork were funded by the Egypt Exploration Society. Additional funding for the IFAO and University of Liverpool fieldwork at Hatnub Project came from IFAO, the British Academy, Fondation Kheops, Gedeon and the BBC.