Thesis subject

MSc thesis topic: National Scale Predictions of Soil Organic Matter: Mechanistic vs. Machine Learning Models

Soil Organic Matter (SOM) plays an important role in climate change mitigation and contributes to soil fertility by enhancing the availability of plant nutrients, improving moisture retention, stabilizing soil structure and increasing permeability among other factors.

National-scale predictions of SOM are essential for land users and policy makers. However, the chosen modelling approach may yield in substantially different estimations of SOM. To test this hypothesis and explore the advantages and disadvantages of different approaches, this research aims to compare the performance of two established approaches currently used in the Netherlands: a statistical and machine learning model (BIS-4D) with a mechanistic model (MITERRA-NL, which is derived from MITERRA-Europe (https://doi.org/10.2134/jeq2008.0108)).

Objectives

  • Predict SOM for the Netherlands using a mechanistic model and a machine learning model
  • Compare performance and assess advantages and disadvantages of the two modelling approaches

Literature

  • Helfenstein, A., Mulder, V.L., Heuvelink, G.B.M., Okx, J.P., 2021. Tier 4 maps of soil pH at 25 m resolution for the Netherlands. Geoderma (accepted & in press)
  • Velthof et al., 2009 (https://doi.org/10.2134/jeq2008.0108)

Requirements

  • (open source) (Q)GIS and R scripting

Theme(s): Sensing & measuring; Integrated Land Monitoring