Dr. Margaret Swift
  • home
  • about
  • cv & publications
  • research =
    • introduction to african savannas
    • ✭ simulating african elephant movements on a fenced landscape ✭
    • mapping waterholes in africa's largest conservation area
    • antelope behaviors on a changing landscape
  • outreach =
    • public talks
    • lesson plans & tutorials
    • teaching statement
    • nsf grfp advice
    • science writing
    • skype a scientist
  • perspectives =
    • those who made me
    • where i live & work
    • decolonization
    • land-grab universities
    • going beyond land acknowledgement
    • asexuality, imposter syndrome, and belonging
    • reading lists
  • art =
    • short stories
    • poetry
    • photography
    • portraits
  • blog
Picture

Simulating African elephant movements
through a One Health lens


In my current postdoctoral project, I use advanced computer modeling to understand how African savanna elephant move about the landscape, and simulate how these patterns might change in the future.

​Watch my 15min conference talk in Kruger National Park HERE!​
​Watch my 40min seminar talk at Cornell HERE!​

Why?

Veterinary cordon fencing in southern Africa has been deployed over many years to mitigate disease transfer from wildlife to livestock, but also limits critical wildlife movements. In the Kavango-Zambezi Transfrontier Conservation Area (KAZA), this fencing prevents elephants (especially breeding herds -- mothers and young) from making their historical migrations. These large-scale movements are important for elephants to access seasonally available food and water. Without these and other resources, elephants and other wildlife may have trouble adapting to ongoing anthropogenic land-use pressures, including losses of water resources driven by climate change.

Recent and growing adoption of non-fence-dependent livestock disease management strategies offers an unprecedented opportunity to re-evaluate fencing policies in KAZA. By simulating elephant migrations, drawing upon real satellite collar telemetry data supplemented by computer modelling and animations, scenarios involving potential fence section removals can be used by policymakers to evaluate the likely impacts of critical land-use decisions.

As ecotourism revenues now rival those of livestock in much of the KAZA region, conservation of large herbivores that shape ecosystems and draw tourists is crucial. In keeping with the One Health theme of the Cornell Atkinson Center for Sustainability, the Cornell K. Lisa Yang Center for Wildlife Health, and the Cornell AHEAD Program (Animal and Human Health for the Environment And Development) this project will highlight the sustainable benefits of optimizing land-use planning at the interface of wildlife, livestock, and human health and livelihoods.

A huge thank you to the Cornell Atkinson Center for Sustainability for funding this project through the Cornell Atkinson postdoctoral fellowship. If you're interested in learning more about my experience at Cornell and the Atkinson Fellowship, please do reach out to me with your questions.

Further reading:​
  • Taylor, R., Naidoo, R., Steel, L., Night, M., Osofsky, S. (2024) "Count, connect, conserve: Southern Africa elephant survey points the way." Mongabay.
  • Osofsky, S.A. (2019) "The global burden of (how we manage) animal disease: Learning Lessons from southern Africa."  Journal of Wildlife Diseases 55 (4).
  • Cumming, D. H. M. et al. (2015) "Beyond Fences: Wildlife, Livestock, and Land Use in Southern Africa." CABI Books.​

How?

I use computer modeling to investigate how elephants currently move around the landscape. These models rely on data from 63 GPS-collared elephants in Namibia and Botswana, and these analyses would be impossible without collaboration from my incredible partners at the World Wildlife Fund and The Ecoexist Project.
  • In Part 1, I use a Hidden Markov Model to characterize activity states (e.g. resting vs. foraging), and a Habitat Selection Model (conditional logistic regression) to determine seasonal landscape use.
  • I use these in Part 2 to simulate elephant movements on the current KAZA landscape via an Agent-Based Model.
  • Finally, in Part 3 I predict how landscape changes may impact elephant distribution by (a) removing some fences (proxying possible fence removal scenarios under actual consideration) and (b) removing ephemeral waterholes (proxying aridification due to climate change), predicting distributional shifts through simulations on the changed landscape.

Picture

The simulations used for this project are built using the framework of the abmAnimalMovement R package. I've extended this code into a separate package to address a few key needs:
  • Introduction of linear barriers (e.g. fences, roads, rivers) with flexibility in perceived permeability
  • Extension of matrix-based landscapes to include real-world georeferenced rasters in simulated agent step choices.
  • Hierarchical shelter use expansion to define both local and global movement restrictions.
  • Seasonal-based selection of landscape rasters and shelter attraction strength in agent movement choices.


Home
About
Contact
​Margaret Swift
Atkinson Postdoctoral Fellow
Cornell K. Lisa Yang Center for Wildlife Health
Cornell University
​Ithaca, New York, USA
  • home
  • about
  • cv & publications
  • research =
    • introduction to african savannas
    • ✭ simulating african elephant movements on a fenced landscape ✭
    • mapping waterholes in africa's largest conservation area
    • antelope behaviors on a changing landscape
  • outreach =
    • public talks
    • lesson plans & tutorials
    • teaching statement
    • nsf grfp advice
    • science writing
    • skype a scientist
  • perspectives =
    • those who made me
    • where i live & work
    • decolonization
    • land-grab universities
    • going beyond land acknowledgement
    • asexuality, imposter syndrome, and belonging
    • reading lists
  • art =
    • short stories
    • poetry
    • photography
    • portraits
  • blog