zoehammond/tanzania_human_elephant_conflict

Spatial analysis of human-elephant conflict risk zones in Tanzania, integrating GBIF occurrence data, WorldPop population density, and WDPA protected areas # Mapping Human-Elephant Conflict Risk in Tanzania A reproducible spatial analysis identifying human-elephant conflict risk zones in Tanzania, integrating species occurrence data, human population density, and protected area boundaries. **Author:** Zoe Hammond, MSc Human Evolution & Behaviour, UCL --- ## The question As a Human Evolution & Behaviour MSc student who is interested in conservation, I wanted to understand where human-wildlife conflict actually happens spatially. Tanzania has both one of Africa's largest African elephant populations and a rapidly growing rural human population. Using only freely available data, can we identify where these populations come into contact, and what does that pattern reveal about how protected areas function? --- ## Headline findings - **168 of 2,000** recorded elephant occurrences (8.4%) fall outside protected areas AND within 10 km of dense human settlement, flagging them as high conflict risk. - **70%** of these high-risk records are concentrated in a single region: **Arusha**. The top three regions (Arusha, Manyara, Mara) account for 92%, corresponding to the well-studied Tarangire-Manyara-Serengeti ecosystem complex. - Conflict-risk records cluster sharply at protected area boundaries. **Median distance to the nearest PA is just 3 km**, and 75% are within 6 km — a textbook *edge effect* signature. --- ## Figures --- ## Methods Three open data sources were integrated: - **Elephant occurrences** — *Loxodonta africana* records in Tanzania from GBIF (n = 2,000) - **Human population density** — WorldPop 2020, 0.5 arc-minute resolution - **Protected areas** — World Database on Protected Areas (WDPA), via the `wdpar` package Analysis was performed in R using the `sf` and `terra` packages. All spatial operations were carried out in UTM Zone 36S (EPSG:32736) for accurate distance calculations. "Dense settlement" was defined as the top quartile of populated cells (>1 person/km²), buffered by 10 km which is app …