Editor’s note: First, since you are currently on the CAAL project website, we hope you will return as the project progresses, the database goes live, and further project outputs are published here. You can see a list of the institutions participating including 2020 reports from 2 institutions digitising their archives, as well as read 53 individual profiles of people actively working on the project. Follow the project @uclcaal on Facebook, Twitter, Instagram or subscribe to the newsletter for monthly updates or just get in touch!
Gai Jorayev introducing the CAAL project
Gai Jorayev introduced the CAAL project where he first discussed the ‘habitation, exchange, and trade of the Silk Road’ and the role of these processes across the geographic region of Central Asia where the remains of tangible and intangible exchanges over vast time periods have left an ‘unparalleled complexity of archaeological sites’. As we have heard in other seminars, interest and research in this legacy has been done over the past 150 years. Much of this data comes from the Soviet era and is held in institutions – mostly in the post-Soviet countries but also around the world. The majority of these archives are in Russian or local languages and accessing this data ‘in order to research the real tangible heritage of Central Asia’ is not an easy task.
‘The main goal of the project is digitising existing documentation of archaeological heritage across Central Asia and we do that not as a small UCL team…but we have extensive partnerships across Central Asia.’ These archives hold reports, photographs, drawings, and monument passports which are specific heritage management records developed during the Soviet period and continued after. ‘Every monument in the system has one of these documents.’ In addition to these forms, identification cards also exist not only for well-known monuments ‘but also for smaller, more ephemeral monuments.’ These are now being digitised in ‘high-quality, high-resolution TIFF files with a lot of metadata’ to which we can add aerial photography and other high-resolution datasets, including ground-level photogrammetric recording with the help of Central Asian colleagues.’
Marco Nebbia then presented the inner workings of how the remote sensing team is working and a case study as a result of this work (see below).
To close the session, Jorayev reiterated the purpose of the project in making these datasets available: current plans in place with the UCL Research Data Storage and the UCL Research Data Repository. These will make static data available while Arches will allow for dynamic use. The project is also ‘in conversation with several major institutions who have been collecting data from the late-1980s to provide links to their data with the idea of creating a single portal where researchers can have an entrance into Central Asian archaeological-related datasets.’ An additional aim is to be able to ground-truth what is turned up through remote sensing. At the moment as you know, travel is restricted, and these plans are on hold but we are looking forward to getting out in the field again.
‘We would like to hand over the tools (through the GIS system) to local institutions, especially the local heritage management institutions so they will [be able] to address their own management challenges using the datasets which were created through the CAAL project.’
The project is ‘raising awareness of the landscapes, challenges, threats the archaeology of the region is facing and hopefully creating some platforms for other areas, such as economy, tourism, infrastructure development that will use the dataset in order to plan activities. Some of this is already happening and we hope to show these case studies in the future.’
Jorayev then thanked other Arcadia-funded projects for collaborating as well as everyone who contributes to the project – in Central Asia and at UCL.
Marco Nebbia on remote sensing
Marco Nebbia began by stressing that his presentation reflected the work of many people organised in different institutions between the UK and Central Asia. These large datasets, like other Arcadia-funded projects, are being handled through Arches open-source platform which allows for storing large amounts of individual records in different languages – we now include English, Russian, Chinese but potentially this can be expanded to other local languages. ‘There are a few limitations, especially to do with its nature of being web-based and the difficulty of adopting this in countries with limited connectivity and it is slightly unstable so multi-editing cannot be handled and it has strong limitations on the analytical capability…. Therefore, we are developing QGIS-based database and using the GeoPackage format which allows for handling different data models, such as raster and various geometries of vector datasets. We are keeping similar resource models used in Arches.’
Currently, archival materials are being imported into the database resulting in 13,325 monuments records, 3,584 archival records from Central Asian partners being combined with 28,082 remote sensing records mapped by the UCL team.
‘In our attempt to replicate the multidimensionality of Arches we are working with different tables and layers which can be of different geometries. For example, we have the archive table with the metadata from digitised documents and monuments records containing information from the passports, actor resource model which lists the institutions working in CAAL and then the remote sensing. To link them all we have a resource relation model which allows us to link all these resources together and feed them into Arches and every archival record will be linked to relevant resources. Crucially, Mahmoud Abdelrazek developed a customised Python script which will allow for multilingual search within QGIS (which is not a native feature of the software).’
The strength of this system is based on the CAAL ID system which organises all the records with unique identifiers within the resource model: heritage, actor, activity, historical event, information, archive, remote sensing polygons, remote sensing lines, remote sensing group, city feature, city. In order to simplify the vast amount of data, smaller units, referred to as regional archaeological mapping teams working on subsets of data to test the linkages and depth of data. This is also a chance for closer collaboration between UCL team members and individual Central Asian partners. It will also allow for the building of regional narratives, that will serve as a compendium to the data collected. The narratives will give a general introduction to each region, and then will describe the landscape, the archaeology and the major threats that sites are exposed to; all supported with visual examples.
The purpose of all this is to make the data useable and sustainable – ‘the effort put into data collection = the effort put into data management and use’. Nebbia stressed that the CAAL project is not only to collect data, but to utilise it.
To these ends, the remote sensing teams have been developing models including: DEM, landcover (crops, urban areas, rivers networks), combined with CORONA imagery and Soviet maps which allow us to identify and track changes over the past 50 years. On top of these are the archaeology (monuments, archives, remote sensing, actors) from which we can test the data sets. In terms of data collection and management, the UCL remote sensing team is scanning the whole area but individuals have been assigned grids so that we can see progress and avoid overlap. We have paid special attention to including an expanded interpretation level as inspired/required/developed by the individuals doing the work who found they needed more space to talk about what they were mapping. Lines and polygons are used as basic layers within the table as well as a polygon group, all linked in a relational table. This is useful in an archaeological landscape which includes individual features such as water courses or burial mounds which exist as single units and groups within a larger landscape. Currently, the teams are using Google Earth, Bing Aerial, and ESRI all streamed into the GIS platform and are freely available. One of the major features of this table is the development of condition assessment and risk assessment for example, to assess the condition, a user can identify the date of the image and then make notes as to the condition of the feature as it can be observed. For risk assessment a series of identifiable threats have been created along with a scale and then the user can include any necessary notations. Ona Vileikis has developed extensive guidelines for the UCL team (these will be available to anyone in future who will be adding to the database). These have been used by Luca Rapisarda in a case study in the Lebap region of Turkmenistan which utilised statistical modelling in conjunction with risk assessment observable through remote sensing.
As mentioned, historical imagery is being used in to observe how the landscape has changed over the past several decades and also allows us to see sites which are no longer visible. An example of this work has been done and written up by Federica Cilio who looked at some specific sites in Khatlon, Tajikistan. Ultimately, all these GIS layers will be combined with the archival data in close collaboration with Central Asian partners to create the complete database.
Nebbia then discussed a spatial risk assessment case study in Khatlon, Tajikistan where a team at UCL have made use of the remote sensing dataset, computational methods and regional risk assessment with a case-control approach in a replicable and customizable model which can be used not only for data in Central Asia but anywhere. This study has been submitted for peer review and is currently in press with the Journal of Cultural Heritage. We will certainly let you know when this is available to read.