Geography
Document Type
Article
Abstract
The populations of African savanna elephants have declined by an estimated 60 % since the 1970s, which can be attributed to a mixture of poaching and habitat loss. Human activities and environmental changes have caused unprecedented loss and fragmentation of elephant natural habitats, resulting in the isolation of elephant populations. Preserving habitat connectivity is thus increasingly important to conserve remaining elephants' populations and maintain ecological functions. A major challenge in large-scale connectivity modeling is data availability constraints. To tackle this issue, we developed an integrated modeling approach that leverages multiple, publicly available occurrence datasets, which vary in format and quality, with a multi-scale SDM to estimate spatial suitability of African savanna elephants. Two SDMs, based on polygon-based observations and presence-only occurrences, were separately calibrated using the Isolation Forest algorithm and then ensembled using Bayes fusion. Particularly, we included multiple landscape metrics derived from a high-resolution (∼5 m) land cover map as environmental predictors in the SDMs to characterize the landscape structure influencing elephant movement. The resulting environmental suitability was then used to map landscape connectivity through circuit theory, implemented in Circuitscape. Using species distribution modeling (SDM) and graph-based landscape connectivity modeling, we aimed to understand population connectivity and target vital corridors across Tanzania, one of the most important elephant range states. Shapley value-based variable analysis in SDM revealed that human modifications strongly influence elephant distribution at broad scales, while habitat fragmentation and connectivity impact their activities. Connectivity results further highlighted that both long- and short-distance connectivity are currently facing significant threats from intensive human activities (e.g., agriculture) in Tanzania and identified critical linkage zones that should be targeted for connectivity conservation efforts. © 2025 The Author(s)
Publication Title
Ecological Modelling
Publication Date
8-2025
Volume
507
ISSN
0304-3800
DOI
10.1016/j.ecolmodel.2025.111198
Keywords
African elephant database, data integration, GBIF, human-wildlife conflict, landscape metrics, Shapley values
Repository Citation
Song, Lei; Frazier, Amy E.; Estes, Anna Bond; and Estes, Lyndon Despard, "A multi-scale approach for integrating species distribution models with landscape connectivity to identify critical linkage zones for African savanna elephants (Loxodonta africana)" (2025). Geography. 1013.
https://commons.clarku.edu/faculty_geography/1013
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.
Copyright Conditions
Song, L., Frazier, A. E., Estes, A. B., & Estes, L. D. (2025). A multi-scale approach for integrating species distribution models with landscape connectivity to identify critical linkage zones for African savanna elephants (Loxodonta africana). Ecological Modelling, 507, 111198.
