From Villages to Empires: Archaeological Settlements of the South Levant from the Chalcolithic to the Byzantine Period

Context

The datasets were collected as part of the project “Governance Policies and Political Landscapes in the Southern Levant under the Neo-Assyrian Empire”, hosted at the University of Turin and funded by the Gerda Henkel Stiftung (grant number: AZ 31/F/21, PI Alessio Palmisano). The project use of a holistic approach integrating archaeological, textual and geographical data into a spatial framework to obtain a better understanding of the Assyrian imperial policies in the Southern Levant region.

It compares radiocarbon dates, archaeological survey and excavation data over the longue durée from two case-study regions, systematically bringing together all available information. This approach mitigates any misinterpretation due to partial use of data and sheds new light not only on the Assyrian governance policies but also on the local responses to the imperial activity.

While the focus of the project is the Iron Age (1150–333 BC), and specifically the Iron Age II (980-539 BC), the dataset includes sites from the Chalcolithic to the Byzantine period. It is well-known, in fact, that in order to understand settlement pattern changes, it is necessary to employ a long-term approach [1, 2, 3], in order to assess the evolution of political landscapes before, during and after the Assyrian domination.

The area under study comprises the two Iron Age II polities of Samaria and Judah, the former a province under the administration of the Assyrian Empire (from 722 BC), and the latter an independent client state. Generally speaking, the interpretation of the Assyrian presence in the area has ranged from the two polar opposites of a destructive entity [4] to an entity capable of bringing prosperity and stability to the area [5, 6]. However, past and recent studies have demonstrated how the Assyrians approach to the management of conquered territories was nuanced, varying between frontier zones and the empire’s core, and more broadly, from region to region [7, 8, 9, 10, 11]. Interpretation of the Assyrian management of the land as “Islands of control” [12] also found good evidence in Samaria and neighboring regions [13, 14]. However, the lack of a systematic study not limited to only excavation data, but that brings together survey data, excavations, and historical sources continues to hamper our ability to understand the settlement dynamics in the region. Furthermore, a multi-scalar approach encompassing neighboring areas and regions [15, 16, 17] helps prevent any misinterpretations by ensuring that broader settlement dynamics typical of the period are not overlooked.

This dataset forms the basis of such a study, offering a highly detailed aggregation of archaeological sites from different published sources and arranged in a coherent spatial database, ideal for both short and long-term analyses (Figure 1).

Figure 1

Spatial coverage

The portion of Southern Levant examined here covers around 14.712 sq.km:

• Description: Palestine and the central regions of Israel

• Geographic Coordinate system: UTM 36 N (EPSG:32636).

• Datum: World Geodetic System (WGS) 1984.

• Northern boundary: 32.57228471 (decimal degrees).

• Southern boundary: 31.13314649

• Eastern boundary: 35.57204756

• Western boundary: 34.47523021

Temporal coverage

4500BC - 638CE

Steps

The dataset was collected in different stages from three main archaeological surveys sources (see Figure 2). First, published archaeological survey maps from the Southern Samaria Survey [18] and Manasseh Hill series of surveys [19, 20, 21, 22, 23, 24, 25] were georeferenced and, where possible, sites were directly digitized on the fly by creating geometries at the point location. However, some publications were lacking a comprehensive survey map (only providing period maps), or did not have enough accuracy in displayed coordinates for the resulting georeferenced process to be considered accurate. In this case, sites were collected in a CSV file with a basic set of information (name, period, size – if available -, bibliographic reference, coordinates pair) and then imported into QGIS and inserted manually into the main point layer. This generally resulted in higher location accuracy than using low-quality georeferenced maps.¹ The sites collected with this method were first digitized using the Old Israel Grid reference system (EPSG: 28193), which is what most survey reports used, and then reprojected in UTM 36 N in QGIS.

Figure 2

The second step was to collect archaeological sites from online databases, in this case specifically the Archaeological Survey of Israel [ASI: https://www.antiquities.org.il/survey/new/default_en.aspx, see 26 pp. 12–15 for more details on the ASI methodology]. This database collects information on archaeological features gathered by surveys carried out in different years and placed online starting more or less 10 years ago. Survey maps in this case are arbitrary squares of 100 km2, and location information are displayed according to the New Israel Grid (EPSG:2039, for technical details on the two reference systems see [27]) or in latitude and longitude pair. Information on site name, bibliographic reference, periods of occupation and location were recorded in different CSVs for each available square in our area of interest and then reprojected in QGIS before including them in our sites layer. A second database used for collecting missing data was the West Bank and East Jerusalem Archaeological Database [WBEJAD, see 26, 28, 29]. In this database, an XLS file with coordinate pairs and details was available and was imported in QGIS as a separate layer before including it into the main sites layer. Both the ASI and the WBEJAD required additional cleaning and harmonization steps before importing them into GIS in a format that was compatible with our sites layer source. These cleaning steps were carried out using the R scripting language and the Positron IDE and will be published together with other analyses in a future publication.

Lastly, the Google Earth Placemark for the Ancient Near East [30] was used to spot missing sites from the above databases, or to fill gaps in non-surveyed areas. Since this dataset does not provide information other than location and site name, other information such as occupational periods and site extension were gathered from published excavation reports.

Each occupational period of a site (site-phase) was recorded as a unique geometry, and the final dataset accounts for 5,542 sites and 14,142 site phases. Site phase here refers to archaeological episodes such as Middle Bronze Age, Iron Age, etc., expressed in approximated calendrical years and indicated by a start and end dates (then convert to Before Present for analyses purposes), see also Table 1 for the periodization adopted. Sub-periodization for each cultural period was also recorded in case more granular chronological information was available from the published material. Since the dataset was used to reconstruct, apart from settlement patterns, also demographic trends by acting as a proxy together with other data, we decided to limit the collection of information to those features which were identified as habitation sites or possible habitations on the original publications. This means that features such as single installations, industrial zones (e.g. mines), and cemeteries, were not included in the collection. Nonetheless, other elements with ephemeral traces of dwellings (such as caves), were still recorded, when they were found to be in use inside the chronological scope of the project.

Sampling strategy

The dataset was gathered from existing publications, but it is worth mentioning that the archaeological sites data derives from excavations and surveys, particularly the latter, that involved a variety of different methods and intensities. While ASI surveys seems to share a similar methodology [26], technologies, techniques, and more importantly pottery sequences knowledge improved during the decades, resulting in uneven data that needed some standardization (see below). For the same reason, multi-decades projects (e.g. Manasseh Hill) have been far more intensive and extensive, and some isolated ASI-sponsored surveys have been more intensive than others. These elements, coupled also with continuous urban sprawling (preventing intensive surveys especially along the coast), resulted in an uneven concentration of sites in the study area. Moreover, except for some isolated hot spots (e.g. the Lachish region), the southern part of our study area is generally less surveyed than the central and northern part (see Figure 3).

Figure 3

Quality Control

All the data in the attribute table have been checked and standardised, when possible, especially regarding cultural periods (see below). These have been standardized according to periodisation adopted, following also previous examples of converting biblical-based periodisations into cultural periods [26]. The location of sites was recorded as georeferenced points from the original coordinate pairs and both the location, and the sites extent have been refined by using multiple sources wherever possible (archaeological survey and excavation reports). Site location quality can be considered relatively high, although we noticed some inconsistencies between the ASI data and the WBEJAD data. Since the latter also involved a site visit and gps recording, we considered it more accurate compared to the former, so we adapted sites present in both sources to the WBEJAD location. To further improve accuracy, we also cross-checked satellite images and published location maps or photograph (when available).

A similar process involved the recording of the sites size. When a site size was readily available, we recorded it (converting it from the source measure to hectares if necessary), and when other type of measurements were given, the area was calculated from those. When a site size was not available or we found conflicting information between different sources, we calculated or adjusted the dimensions using remotely sensed images. Since each site phase represent a site occupational period, when available we recorded different sizes for different periods. This was already provided by some source survey data [18], from intensive surveys on excavated sites [e.g. 33], from specific site studies [34], or from archaeo-demographic studies [35, 36, 37].

Constraints

Generally speaking, site dimensions are provided in many different ways across different archaeological projects and will always involve a degree of inaccuracy. Manually calculating an area based on site measurement (e.g. width and length) can lead to artificially inflating site sizes, but it has been proved that for smaller sites, this inflation is negligible [38]. The period-based size will always be an approximation, but in case of excavations and intensive surveys, we consider it a better solution than assigning the whole site extent to every period. Both the location accuracy and the size accuracy were represented with a quality scale (from A to D) in the attribute table, based on the quality of the source data (see the sites.txt file for a description).

The start and end dates of the cultural periods represent absolute calendric years for each site phase. These dates are based on typo-chronological schemes defined by previous excavations and specialist studies. Of course, there will always be a varying degree of uncertainty in dating site-phases and periods based on cultural material (e.g. pottery), which can result in very short site phases (spanning a few decades) or larger ones (spanning centuries). This is especially true for survey data, where occupational periods are sometimes harder to refine due to the presence of long-lived pottery types and the lack of stratified materials or radiocarbon dates from the site itself, or accurate anchor points from reference materials of excavated sites in the region. The result of this uncertainty is that certain site duration can span several centuries depending on the dating methods adopted (see e.g. Figure 4).

Figure 4

Object name

• dataset – a geopackage database containing two layers:

  • archaeo_surveys_area – a polygon layer representing the spatial extent of the archaeological surveys from which se-ttlements data have been collected.

  • archaeo_sites – a point layer of archaeological sites in South Levant.

• archaeo_sites – a set of two files respectively providing a spreadsheet (.csv) of sites in South Levant, and a field description for the attributes table of the sites (.txt).

Data type

Secondary and Processed data from originally published materials.

Format names and versions

.csv, .txt, .gpkg

Creation dates

The dataset was created in 2023–2025 as part of the Gerda Henkelf Stiftung funded “Governance Policies and Political Landscapes in the Southern Levant under the Neo-Assyrian Empire” project.

Dataset Creators

The researcher responsible for the creation of the dataset and management was Andrea Titolo, data entry was also carried out by Alessio Palmisano. Database design was carried out by both authors.

Language

English

License

CC-BY 4.0: https://creativecommons.org/licenses/by/4.0/.

Repository location

The dataset is deposited at the Zenodo repository under DOI: https://doi.org/10.5281/zenodo.15111732.

Github location

https://github.com/UnitoAssyrianGovernance/villages-to-empire-dataset.

Publication date

2025-03-31