Thesis subject

MSc Thesis topic: Mapping habitat suitability in the Netherlands with Machine Learning

Wageningen University & Research develops and maintains the Model for Nature Policy (MNP), which is one of the key tools available to policy makers for evaluating the effects policy will have on biodiversity. It is actively used by the Netherlands Environmental Assessment Agency (PBL) in many different projects.

The Model for Nature Policy (MNP) makes predictions on a species' viability within the Netherlands, through Habitat Suitability (HS), using expert judgement and environmental factors. MNP predictions are validated by visually and statistically comparing HS maps generated by MNP to data from the Dutch National Database Flora & Fauna (NDFF) and the Vegetation of the Netherlands (DVN). The NDFF data set only containing information on a species' presence, offering no information on its absence. Nowadays, the incorporation of crowed-sourced data (as those of citizen science projects) may lead to rare species to be over represented, as a result of them being valuable 'trophies' to citizen scientists. The proposed student project involves using NDFF and DVN data sets, environmental factors sourced from the Basisbestand Natuur en Landschap (BNL) to explore machine learning as a means to improve the scientific basis of MNP, for example by making HS maps from presence data for use during validation.

Objectives

  • Get familiarized with Machine Learning methods for species distribution modelling.
  • Produce Habitat Suitability maps for a selection of species and compare them to currently used maps.

Literature

Requirements

  • Machine Learning (FTE-35306) or Deep Learning (GRS-34806)
  • Programming in Python, Geo-Scripting

Theme(s): Modelling & visualisation