Student information

MSc thesis topic: Improving Bird Species Distribution (and Abundance) Models By Integrating Uncertainty

Species distribution models (SDMs) are commonly used to describe and predict the spatial distribution and abundance of species. Observational data, collected at a part of the study area, is related to environmental data to predict the possible presence and absence of a species as well as their abundance at unvisited locations.

At Sovon, the Dutch Centre for Field Ornithology, SDMs are already successfully used in predicting the spatial distribution of bird species and making nation-wide distribution and abundance maps (Figure 1). The necessary observational data as input for these models consists of both presence and absence, as well as the abundance of a species at a certain moment of the year. The observations is done by a variety of volunteers, with different strategies and approaches. To keep improving their SDMs Sovon likes to explore methods to include the uncertainty or reliability of the observational data and its impact on the SDMs. Several metadata and attributes related to the observations potentially hold valuable information on the reliability or certainty of each observation. Sovon would like to include this indication of uncertainty in their SDM by prioritizing more reliable observations to obtain better estimates of the distribution and abundance of bird species in the Netherlands.

Relevance to research/projects at GRS or other groups

It is hypothesized that integrating known observation reliability or uncertainty will lead to improved SDMs and better predictions of spatial distribution.

Exact Research Objectives will be formulated with supervisor, but potential objectives are:

  • Review literature on methods to integrate uncertainty of observational data into Species Distribution Models.
  • Develop method to integrate uncertainty in SDM and assess the impact of the resulting predictions.

Literature

  • Beale, C. M., & Lennon, J. J. (2012). Incorporating uncertainty in predictive species distribution modelling. Philosophical Transactions of the Royal Society B: Biological Sciences, 367(1586), 247-258.
  • Tessarolo, G., Ladle, R. J., Lobo, J. M., Rangel, T. F., & Hortal, J. (2021). Using maps of biogeographical ignorance to reveal the uncertainty in distributional data hidden in species distribution models. Ecography, 44(12), 1743-1755.

Requirements

  • Passion for birds and open for a data-driven perspective on bird species distributions
  • Knowledge of R is desirable

Theme(s): Modelling & visualisation