Thesis subject

MSc thesis topic: Incorporating viewsheds in distance samplingtechniques to estimate wildlife populations in the Kgalagadi Transfrontier Park

Estimating population sizes of wild animals is crucial for nature conservation (Davis & Winstead, 1980; Norton‐Griffiths, 1978; Van Lavieren, 1982). Most wildlife population estimates are based on sample counts along transects, where animals are only counted in a part of the area whereafter this number is extrapolated to cover the entire area. This counting method is cheaper and less time‐consuming than total counts (Jachmann, 2001; Norton‐Griffiths, 1978; Van Lavieren, 1982).

The present‐day method of estimating wildlife populations from count data has not changed substantially since the 1970s, and suffers from imprecise and inaccurate population estimates (Eikelboom et al., 2019). Modern GIS techniques and the use of remote sensing data can potentially improve upon some convenience-based statistical assumptions for the extrapolation of sample counts to population estimates.

Background

Distance sampling is a widespread technique to estimate wildlife populations, where the distance to each counted animal is measured and used to extrapolate the counts to total population estimates using detection functions that approximate the diminishing animal detection probability over distance (Buckland et al. 2004; 2015). However, this approach assumes that the landscape is homogeneous in terms of visibility, which in practice is not the case due to fluctuations in the terrain and landscape elements that obstruct line of sight (e.g. vegetation). Modelling viewsheds from the perspective of the observer along the survey transect could allow us to explicitly take the heterogeneity of the landscape in terms of visibility into account, which can be incorporated into the statistical techniques to extrapolate the count data to population estimates.

The study area for this project is the Kgalagadi Transfrontier Park, a vast wildlife reserve spanning the border between South Africa and Botswana. This park is characterised by its unique ecosystem, which includes semi-arid grasslands and desert, interspersed with sporadic vegetation and undulating dunes. These features make it an ideal study location to apply and test the effectiveness of integrating viewshed analysis into wildlife population estimation.

Relevance to research/projects at GRS or other groups

For this project, Maya Beukes from Senckenberg Research Institute (the researcher who collected all data) is also available as an external advisor

Objectives and Research questions

  • Develop a vegetation map based on satellite imagery for the Kgalagadi Transfrontier Park and translate this into a digital surface model (DSM) using average plant species height and shape for this region
  • Develop a model to simulate a viewshed for all points along the survey trajectories in the Kgalagadi Transfrontier Park using a DSM based on the Digital Elevation Model and vegetation surface model
  • Develop a distance sampling statistical model that combines visibility from the viewshed model with the detection function to estimate the absolute wildlife population sizes in the Kgalagadi Transfrontier Park for the riverbeds and dunes strata separately
  • Evaluate the loss in precision and accuracy of the estimated population sizes when applying a modified version of this model to wildlife count data for which the distance and direction to the observer is unknown

Requirements

  • GIS and remote sensing knowledge, to classify vegetation and construct a DSM
  • Experience with statistics by having completed e.g. Advanced Statistics (MAT20306) or Ecological Methods (WEC31806)
  • Programming experience (R or Python)

Literature and information

  • Davis, D. E., & Winstead, R. L. (1980). Estimating the numbers of wildlife populations. In S. D. Schemnitz (Eds.), Wildlife management techniques manual (4th ed., pp. 221–245). Washington, D.C.: The Wildlife Society.
  • Norton‐Griffiths, M. (1978). Counting animals: Revised second edition. Handbook No. 1. Serengeti Ecological Monitoring Programme. Nairobi, Kenya: African Wildlife Leadership Foundation.
  • Van Lavieren, L. P. (1982). Wildlife management in the tropics ‐ with special emphasis on South East Asia: A Guidebook to the Warden. Part 1 ‐ Introduction, Taking Fieldnotes & Wildlife Census Methods. Ciawi ‐ Bogor, Indonesia: School of Environmental Conservation Management.
  • Jachmann, H. (2002). Comparison of aerial counts with ground counts for large African herbivores. Journal of Applied Ecology, 39, 841–852.
  • Eikelboom J.A.J., Wind J., Van de Ven E., Kenana L.M., Schroder B., De Knegt H.J., Van Langevelde F. & Prins H.H.T. (2019). Improving the precision and accuracy of animal population estimates with aerial image object detection. Methods in Ecology and Evolution 10(11): 1875-1887.
  • Buckland, S.T., Rexstad, E., Marques, T.A. and Oedekoven, C.S. 2015. Distance Sampling: Methods and Applications. Springer, Heidelberg.
  • Buckland, S.T., Anderson, D.R., Burnham, K.P., Laake, J.L., Borchers, D.L. and Thomas, L. (Editors) 2004. Advanced Distance Sampling. Oxford University Press, Oxford, UK.

Theme(s): Sensing & measuring; Modelling & visualisation