The Sudan Middle Stone Age Lithic and Contextual Database (SUM): Structure, Research Potential, and Future Directions

Context

Sudan occupies a strategic position in northeastern Africa, linking sub-Saharan Africa, Northern Africa, and the Red Sea regions, while also offering a potential route for expansion through the Nile Valley and Red Sea corridors toward the Levant and Arabia [1, 2, 3, 4, 5]. This location has long been recognized, in many studies, as central to understanding population movements, the transmission of technological traditions, and interactions with different ecological niches [2, 5, 6]. Consequently, the Sudanese MSA provides a crucial comparative framework for examining broader questions of modern human dispersal, cultural innovation, and adaptive strategies in arid, riverine, and coastal environments [2, 4, 7, 8, 9, 10, 11, 12, 20, 21, 22, 23, 25]. Despite this importance, the Sudanese archaeological record remains fragmented: known MSA sites are widely but unevenly scattered across the country (Figure 1), and many are documented only in preliminary reports or gray literature, with limited stratigraphic detail, inconsistent or nonstandard terminology, and variable technological descriptions. Chronological control is particularly weak, as only 17 of the 283 known sites are absolutely dated. Moreover, the use of different dating methods and their uneven distribution across regions make cross-regional comparison difficult. Site integrity is also often compromised by the dominance of surface collections, with relatively few sites benefiting from systematic excavation or geoarchaeological investigation [12, 13, 14, 15, 16].

Figure 1

Against this background, this study aims to make the Sudanese MSA record systematically accessible for future research on population movement and technological transmission. To this end, SUM has been developed as a clean, standardized dataset that integrates geographical, chronological, environmental, and technological information from 283 sites, including regional lithic technological diversity – such as the Nubian Levallois reduction method, raw material procurement strategies, and blade production – together with the environmental and chronological contexts in which these technologies occur. Importantly, SUM includes data from both published studies and the grey literature (see Figure 2), making a substantial portion of the Pleistocene archaeological record of Sudan accessible for the first time. Through a critical assessment of available data and the application of controlled terminology, the database ensures the comparability of different cases and allows for a transparent evaluation of technological diversity and contextual integrity. While recent synthetic contributions have advanced our understanding of the Sudanese and northeastern African MSA record – notably Usai [7], Masojć [17], Leplongeon [9], and Garcea [10]—these works necessarily rely on the published literature and are not structured around systematic primary data. SUM complements and extends these syntheses by providing the first openly accessible, standardized raw dataset for the Sudanese MSA, enabling future researchers to interrogate the primary evidence directly. In doing so, it lays the groundwork for situating Sudan within current debates on Nubian Levallois technology, blade production, raw material strategies, and their implications for models of human dispersal and ecological adaptation.

Figure 2

Dataset Aims

The main aim of this database is to make information on Sudanese MSA sites freely accessible in a structured and reusable format. By integrating spatial, chronological, raw material, and technological data, the dataset provides a comprehensive basis for comparative research at local, regional, and inter-regional scales. A key goal is to ensure transparency by preserving original raw data alongside standardised categories, allowing users to consult primary descriptors and independently verify analytical decisions. The database is designed to support a wide range of applications, including spatial modelling, technological synthesis, chronological frameworks, and assemblage comparisons. By making this resource openly available, it is intended that future research will provide robust evidence for understanding Sudan’s role in debates on human dispersal, technological diversity, and ecological adaptation during the Middle and Late Pleistocene.

Spatial coverage

This dataset presents the distribution of Middle Stone Age sites in Sudan recorded across published and unpublished sources, including peer-reviewed literature, grey literature, and ongoing fieldwork data.

Northern boundary: +22.053775 (Wadi Halfa 34A)

Southern boundary: +12.000000 (Nyertete)

Eastern boundary: +38.040000 (Grar-South 02, Red Sea Region)

Western boundary: +24.100000 (Nyertete)

Nyertete marks both the southern and western limits of the database. Of the 283 entries, 263 have high-accuracy coordinates, while 20 – concentrated among the Sai Island cluster (Between the Second and Third Cataracts), Grar-South 01 and 02 Red Sea Region [26, 42], BP424 and BP1316 Bayuda Desert [41], 3-Q-34, Wadi Umm Rahau HP732, and Hagar Sail HP15 (Fourth Cataract Region) [39, 43].

Temporal coverage

Many of the Sudanese MSA sites are surface occurrences and therefore lack absolute chronology. Only 17 radiometric ages are currently associated with sites in the database, obtained using OSL, U-series, U-Th, and radiocarbon methods. These ages range from a possible MIS 7 context at Sai 8-B-11 (c. 223 ± 19 ka; [12]) to a MSA occupation at HP766 (~21,450 ± 140 uncal BP; [18]). Several of these dates derive from unpublished or forthcoming data, including two sites in the Eastern Kerma Reach (SK7, SK87; ~50 ka [44]), one in the Wadi Halfa area (ANW-3; ~65–62.5 ka [22, 23]; [6, 19]), one in the Southern Dongola Reach (Affad 23; ~58–50 ka [15, 25]), and two in the Bayuda Desert (BP177; 62–17 ka [24], BP424; 59.4 ± 7.3 ka [41]), with further chronological work ongoing at additional Bayuda sites (e.g., BP1316 [41]); as these analyses are formally published, the dataset’s chronological resolution is expected to improve further (see Table 1 for details).

Sampling strategy

The sampling strategy for SUM was intentionally inclusive: all Sudanese sites attributed in the bibliography to the MSA, Middle Palaeolithic, or Late Pleistocene were included, with the only condition that a location could be obtained, either from reported coordinates (flagged as high accuracy) or by georeferencing a published or unpublished site map (flagged as moderate accuracy). Full technological descriptions and raw material data were not required; sites with partial, uncertain, or absent information were retained, coded 0 for explicit absence and NA for unreported attributes, to preserve a transparent record of the evidence and its limitations. Table 2 summarizes the inclusion/exclusion logic applied to each record.

Steps

The SUM Database was developed in four stages: literature collection, data extraction, data organization and data cleaning.

Limitations

As a database, SUM has several limitations that users should consider when interpreting its contents, especially when drawing regional or temporal conclusions. These limitations are summarized below, together with explicit guidance on how each should inform interpretation of the dataset.

Chronological coverage. Only 17 of 283 sites are absolutely dated, and dating resolution varies; most dated sites are rated Reliable, though 4 are Partially Reliable and 1 is Not Reliable. Guidance: consult the Reliability of Age Estimates field before treating any date as secure, and avoid building regional or temporal narratives on undated or Partially Reliable entries – apparent chronological gaps or clusters may reflect research coverage rather than past occupation patterns.

Grey literature. A substantial share of records derives from unpublished theses, project reports, and archival sources that have not undergone peer review. Guidance: these entries are flagged; users should treat uneven data density across regions as a potential artifact of research and publication intensity rather than genuine settlement variation.

Collection method. Many sites are represented only by surface collections, which lack secure stratigraphic and chronological context and may mix multiple occupation phases. Guidance: users conducting spatial or technological analyses should filter by collection method and treat surface-derived patterns with caution, as they may reflect preservation or collection bias rather than past behaviour.

Data structure

The main database includes various data types, such as spatial coordinates, age ranges, chronological data, raw material records, categorical classifications, and technological attributes (Table 2). It is organized across five sheets: SH1 preserves the original raw data as recorded in the source publications; SH2 contains the standardized, clean data for all variable categories; SH3 provides column definitions and controlled vocabulary; SH4 and SH5 contains the full bibliography. Technological data are organized into two sheets (SH1 and SH2). SH1 records the original terminology used by the authors, while SH2 presents standardized categories, allowing direct comparison between the two. Nubian I, Nubian II, and Aterian points are retained as independent categories in both sheets because of their significance as cultural and dispersal markers. Where data is not available or relevant for a site record, the field is left blank or recorded as NA.

Site location and spatial accuracy

Site coordinates (latitude and longitude) were taken directly from published literature where available and converted to decimal degrees. When coordinate data were not provided, published or unpublished maps and figures were georeferenced against Google Earth satellite imagery to determine site locations [27, 28]. The quality of location data was classified into two categories (Table 4): High, where coordinates were directly reported by the original authors and correspond to the documented site location, suitable for local- and landscape-scale analyses; and Moderate, where coordinates were derived by georeferencing a published or unpublished site map rather than taken from directly reported values, and are therefore best suited to broader regional summaries [26, 42]. When coordinates were identified as erroneous – for example, transposed or clearly misplaced values – the record was corrected where possible or flagged accordingly, with a note documenting the nature of the discrepancy. For each entry, the source used to verify the location and a brief justification for the assigned category are recorded in the database.

Site’s details

The database records 276 open-air sites and 7 rockshelters, most represented exclusively by surface material (Table 5). The best-represented regions are the Eastern Kerma Reach (n = 91) and Southern Dongola Reach (n = 64), followed by the Wadi Halfa Reach and Northeastern Sahara (n = 43), Between the Second and Third Cataracts (n = 16), Third Cataract Region (n = 14), and Northwestern Sudan (n = 13). Other regions are represented by only one to seven sites each. This distribution reflects the historical concentration of archaeological research across Sudan rather than the actual density of MSA occupation.

Data sources

The main database includes various data types, such as spatial coordinates, age ranges, chronological data, raw material records, categorical classifications, and technological attributes (Table 3). The database is organized into five sheets. SH1 preserves the original raw data exactly as recorded in the source publications. SH2 contains standardized, cleaned data for all variable categories, used for comparative analysis. SH3 provides the column definitions and controlled vocabulary underlying the standardized categories in SH2. SH4 and SH5 contain the full bibliography, covering the literature used to build the controlled vocabulary and the literature used to compile the site-level data, respectively. Technological data specifically are organized across two sheets (SH1 and SH2), allowing direct comparison between original and standardized terminology. All published sources are listed in full across these sheets, and each site record is individually linked to its primary reference field in SH1 and SH2, so that entries can be traced to their source without cross-referencing SH5. The authors welcome corrections or additions for future versions of the database.

Data Availability

The Sudanese MSA database is deposited on OSF (https://osf.io/z3t62/overview), available in Excel and CSV formats, and licensed under CC BY 4.0. It includes a standardized dataset component for comparative use, accompanied by a terminology and definitions sheet that serves as a codebook/README explaining categories and coding rules. The dataset supports a range of concrete applications. It enables spatial analyses of site distribution, such as mapping the density and clustering of MSA occupations across regions or testing associations between site locations and landscape features (e.g., proximity to relict water sources). It supports technological comparisons between regions, such as contrasting the frequency of Nubian Levallois reduction or Sangoan large cutting tools across the Nile Valley, Bayuda Desert, and Red Sea regions. The dedicated reliability fields allow users to assess or model chronological confidence, for example by weighting or filtering absolute dates according to their assigned reliability rating before constructing regional chronological frameworks. Finally, the standardized spatial and chronological data can be integrated with palaeoclimatic and dispersal-corridor datasets to examine the timing and environmental context of proposed Nile Valley and Red Sea dispersal routes. The dataset is openly licensed for reuse and integration into other resources, provided appropriate acknowledgment is given.

Object name

SUM. Sudan Middle Stone Age Lithic and Contextual Database

SH1. Raw data

SH2. Main database (clean data)

SH3. Column definitions and controlled vocabulary

SH4. Bibliography used to build the vocabulary, see SH3

SH5. Bibliography used to build the Sudanese MSA information, see SH1 and SH2.

Data type

Secondary data compiled from previously published and unpublished (grey) literature, including peer-reviewed publications, unpublished theses, project reports, archival records, and personal communication with active researchers.

Format names and versions

.xlsx and csv

Creation dates

Systematic compilation of the dataset commenced in 2023, with updates throughout 2024 and 2025, culminating in its current version in 2026.

Dataset Creators

Osman Khaleel, Sol Sánchez-Dehesa and Nuno Bicho. All published literature and references consulted in the compilation are provided in full in SH4 and SH5.

Language

English

License

CC-By Attribution 4.0 International

Repository location

The SUM database is permanently archived on the Open Science Framework (OSF) at https://osf.io/z3t62/overview under the title “The Sudan Middle Stone Age Lithic and Contextual Database (SUM): Structure, Research Potential, and Future Directions,” registered under Universidade do Algarve, Portugal (UALG). The dataset is distributed under a CC BY 4.0 Attribution International license in .xlsx and .csv formats, and is accompanied by a README file and a codebook (SH3: Column definitions and controlled vocabulary) documenting all fields, coding conventions, and controlled terminology.