(1) Overview
Repository location
Context
Across tropical forests, both biodiversity and the local ecological knowledge that helps people recognize and manage it are increasingly threatened, making it important to document and conserve this intertwined natural and cultural heritage (Díaz et al., 2019; Fernández-Llamazares et al., 2021). Local ecological knowledge constitutes an important dimension of biocultural diversity because it reflects long-term interactions between human communities and their environments through language, classification systems, and ecological practices. Documenting it is particularly important in tropical forest regions, where both biodiversity and culturally embedded knowledge systems are increasingly affected by environmental change, globalization, and sociolinguistic transformation (Maffi, 2005; Berkes, 2018). In this context, datasets combining biodiversity-related knowledge with linguistic information remain relatively scarce for many African forest regions, including western Uganda. In addition to its conservation relevance, this dataset captures culturally embedded knowledge systems through the documentation of vernacular taxonomies and naming practices in Rutooro. Such data provide insights into how language encodes ecological knowledge, how categories of living beings are culturally structured, and how knowledge is shared within and across communities. The dataset thus contributes to broader debates in linguistic anthropology, ethnobiology, and the study of human–environment interactions. We provide a two-step survey dataset collected from research and conservation-project staff working in Sebitoli, the northern part of Kibale National Park (Uganda) to document local knowledge of tropical forest wildlife. The Sebitoli Chimpanzee Project (SCP) was established in 2008 by the Muséum National d’Histoire Naturelle (MNHN) and the NGO Great Ape Conservation Project. It employs people from nearby villages who are of Tooro culture. The staff consists of several teams carrying out research and biodiversity monitoring tasks through remote sensing, transect maintenance, anti-poaching patrols, awareness-raising, and management activities. The study area covers 25 km² of forest located in southwestern Uganda, in the districts of Kabarole and Kyenjojo. All data were collected with the approval of Uganda Wildlife Authority, Uganda National Council for Science and Technology in Uganda, and Makerere University (Kibale data, research permit COD/96/05).
(2) Method
Steps
The survey took place in two successive steps. In Step 1, participants were presented with a visual cue (photo and/or video) for each focal taxon and asked to provide an English name and a local (Rutooro) name. For a subset of taxa, participants also listened to an audio recording presented separately and provided English and local names. In Step 2, participants were given all available cues (name, photo and/or video, and sound where available) and rated how frequently they see and (when relevant) hear each taxon on a 0–3 ordinal scale (0 = never, 1 = once a year, 2 = many times in a year, 3 = every week) in the forest area.
Sampling strategy
The survey was designed by the MNHN research team and administered to research-station staff as part of their routine work to produce data supporting the SCP’s research and conservation activities. In total, 34 participants from SCP team took part and were interviewed in group sessions, while responses were recorded at the individual level. Participants provided informed consent. The dataset contains anonymized participant codes and no directly identifying personal information.
The survey covers 54 vertebrate taxa from African tropical forests (32 mammals, 15 birds, and 7 reptiles). The initial species list was derived from published work in the same area (Wanyama et al., 2010), and was then expanded to include additional taxa based on the conservation concerns (threatened species listed on International Union for Conservation of Nature (IUCN) red list, species impacted by road kills (Krief et al., 2020)) and research topics (birds as biodiversity indicators), and to balance the biodiversity representation. In Step 2, the survey was designed as a closed-list recognition task to ensure standardized comparisons across participants and taxa; participants were therefore asked to report observation or acoustic detection frequencies only for species included in the predefined list. We present an example of a visual cue in Figure 1.
Figure 1
Frame extracted from the chimpanzee (Pan troglodytes) video cue. ©SCP.
The 54 taxa included in the survey were not intended to provide a comprehensive inventory of the local vertebrate fauna, but rather a curated subset representing key functional and conservation groups within the study area. Although a more open-ended design could have captured a broader range of species observations, the study aimed to assess perception and recognition patterns across a fixed set of ecologically relevant taxa rather than produce an exhaustive species inventory.
Quality control
Data processing focused on standardizing textual responses while preserving participants’ original terminology. English common names were corrected for spelling and harmonized when obvious variants referred to the same taxon. For local names recorded in Rutooro, orthographic variants were standardized when they clearly referred to the same lexical root. This included harmonizing common noun structure (e.g., Eki-, Ki-, Ebi-, Bi-, corresponding to singular and plural, or forms without prefix) and consolidating phonetically similar spellings that reflected the same term (e.g., Eriba, Eliba, Eriiba (en: Dove)) (Kaji, 2009). Such variation in transcription or everyday usage can be expected during this type of survey. In such cases, standard noun structures were selected without prefix, in the singular form and based on the spelling suggested by Rubongoya (1999). No other transformations or inferential processing were applied.
(3) Dataset Description
The dataset contains four files. Participants_information.csv provides background information for the 34 participating SCP staff members, including their code, their SCP work unit (anti-poaching, camp, chimps guardians, field assistants, sensitisation, and trail maintenance), their working schedule (days or days & nights), the approximate number of hours spent in the forest per week, year of recruitment at SCP, and local tribe. The participants are involved in routine research, monitoring, and conservation activities. Other information was included because it may influence participants’ exposure to wildlife, familiarity with taxa, and local ecological knowledge. For example, work unit, working schedule, and time spent in the forest may affect encounter opportunities, while recruitment year may reflect field experience. Local tribe information was included because vernacular naming practices and ecological knowledge may vary across cultural and linguistic backgrounds. Most participants (32 out of 34) belonged to the Tooro community, which is the predominant ethnic group in the area.
Scientific_names.csv lists all 54 taxa used for the survey, with English names, Rutooro names (when existing), scientific names and taxonomic level. Sebitoli_survey.xlsx contains the survey responses, in three sheets: the Step 1 – Names (ORIGINAL) worksheet stores the raw taxon identification responses (English and local/Rutooro) from the 34 participants for 54 image/video items and 26 sound-only items before formatting and standardization; the Step 1 – Names (STANDARDIZED) worksheet contains the standardized version of these names used for analyses, based on the standardization process described in the section above; the Step 2 – Frequency worksheet stores encounter-frequency ratings (see and hear; 0–3 scale) from the participants for all taxa. Finally, the file image_and_sound_cues.zip contains the 54 images and 26 audio files used as stimuli during the survey. All materials were obtained from the SCP database or, when the target species was not available in the SCP database, from other publicly accessible sources listed in the Sources.csv file included in the archive.
To provide a rapid overview of encounter patterns, we present a summary table reporting the 10 taxa most frequently reported as seen and heard, based on the mean frequency scores averaged across all participants (Table 1).
Table 1
Top 10 taxa most frequently reported as seen and heard (mean across participants).
| TAXON | MEAN SEEN FREQUENCY SCORE |
|---|---|
| Olive baboon | 2.97 |
| Black and white casqued hornbill | 2.82 |
| Elephant | 2.62 |
| Chimpanzee | 2.56 |
| Great blue turaco | 2.53 |
| Vieillot’s black weaver | 2.44 |
| Grey crowned crane | 2.44 |
| Uganda blue headed tree agama | 2.38 |
| Red eye dove | 2.35 |
| Hadada ibis | 2.35 |
| TAXON | MEAN HEARD FREQUENCY SCORE |
| Black and white casqued hornbill | 2.94 |
| Vieillot’s black weaver | 2.73 |
| Red eye dove | 2.68 |
| Great blue turaco | 2.67 |
| Grey crowned crane | 2.59 |
| Chimpanzee | 2.59 |
| Olive baboon | 2.56 |
| Hadada ibis | 2.56 |
| Elephant | 2.35 |
| African emerald cuckoo | 2.29 |
Repository name
Open Science Framework (OSF)
Object name
Dataset on a research project staff knowledge of local taxa and encounter frequencies in Kibale National Park, Uganda
Format names and versions
CSV, EXCEL
Creation dates
2025-11-13 to 2026-02-25
Dataset creators
Gabriel Dubus, Hugo Magaldi, Raymond Katumba, Harold Rugonge, John Justice Tibesigwa, Marc Allassonnière-Tang, Sabrina Krief, and the SCP staff who participated in the survey.
Language
English, Rutooro.
License
CC Attribution 4.0.
Publication date
2026-03-13.
(4) Reuse Potential
The dataset can be reused in fields such as linguistics, anthropology, and ethnobiology, among others. The inclusion of vernacular (Rutooro) names enables analyses of taxonomies, lexical variation, and the relationship between language and ecological knowledge. For example, the dataset can be used to investigate how different taxa are categorized, named, and cognitively organized, as well as how such knowledge varies across individuals and social groups within a shared cultural context. Such questions have been central in ethnobiological and linguistic anthropological research examining how local communities classify living beings and encode ecological knowledge through language (e.g., Berlin, 1992; Atran & Medin, 2008). This knowledge is also relevant for studying the interaction between humans and the environment they live in (Folke, 2006; Berkes, 2018). In a broader picture, the dataset supports interdisciplinary research bridging human sciences and natural sciences, contributing to the documentation and preservation of biocultural diversity. Furthermore, it also provides material for studying local naming practices and culturally situated ecological knowledge in western Uganda.
Beyond these humanities-oriented applications, the dataset also provides a baseline assessment of how well research and conservation staff recognize a set of common tropical forest vertebrates using visual cues and auditory cues, and how frequently they encountered each taxon. It can be reused to (i) compare human knowledge with biodiversity metrics derived from other survey approaches (e.g., camera traps and passive acoustic monitoring) in the same landscape, (ii) benchmark and improve field protocols, for instance by identifying taxa that are frequently misidentified and guiding targeted training, and (iii) support descriptive and comparative analyses of recognition performance across taxa, cue types, or staff profiles (e.g., team, years of experience or time spent in the forest), where such comparisons are appropriate and ethically justified. The dataset also includes vernacular (Rutooro) names where available, which may be useful for ethnozoological research, community engagement materials, and the development of multilingual field guides or training resources.
Documenting vernacular taxonomies and ecological knowledge contributes to safeguarding the biocultural heritage of Rutooro-speaking communities in western Uganda, where cultural knowledge and natural heritage are deeply interconnected. At a time when tropical forest loss is eroding both local knowledge systems and the habitats of emblematic and threatened species, such documentation can support the preservation of community knowledge alongside biodiversity conservation.
Potential limitations and barriers to reuse include the fact that identification responses are free-text and may contain spelling variants. Although basic normalization and scoring approaches can be applied, alternative scoring rules may yield slightly different accuracy estimates. In addition, local name “accuracy” depends on the availability of a reference local name for each taxon. Frequency ratings are self-reported ordinal measures and may reflect perceived encounter rates rather than standardized effort, so they should be interpreted as relative indices rather than direct measures of abundance. In addition, observation dates for the different species recorded during the survey were not reported in the dataset. Moreover, the 54 taxa included in the survey were intentionally selected as a representative subset of ecologically and conservation-relevant vertebrates rather than as a comprehensive inventory of the local fauna. Finally, because the dataset involves human participants, reuse should respect the dataset’s anonymization and consent conditions and avoid attempts at re-identification or sensitive profiling of individuals.
Acknowledgements
We are grateful to the Ugandan Wildlife Authority and the Ugandan National Council for Science and Technology for permission to conduct this research at Sebitoli. We deeply thank Jean-Michel Krief, co-director of the Great Ape Conservation Project, as well as the staff members of the Sebitoli Chimpanzee Project (SCP) who participated in this project.
Author Contributions
Gabriel Dubus and Hugo Magaldi contributed equally and share first authorship.
Gabriel Dubus: Data curation; Methodology; Writing – Original draft.
Hugo Magaldi: Data curation; Methodology; Writing – Original draft.
Raymond Katumba: Investigation; Methodology.
Harold Rugonge: Investigation; Methodology.
John Justice Tibesigwa: Conceptualization.
Marc Allassonnière-Tang: Methodology; Writing – Review & Editing; Supervision.
Sabrina Krief: Conceptualization; Methodology; Writing – Review & Editing; Supervision; Funding acquisition.
All authors reviewed and approved the final manuscript.