MASTER Assignment
Soil organic carbon mapping for farms in the netherlands
Type : Master M-BIT
Period: Oct, 2022 - Apr, 2023
Student : Giessen, D.R. van de (David, Student M-BIT)
Date Final project: April 26, 2023
Supervisors:
Abstract:
The aim of this study is to develop a method to predict the soil organic carbon contact on farm plots in the Netherlands. It first provides a comprehensive systematic literature review to understand the state of the art and best practices used in similar studies. The results of this literature review include a set of environmental covariates, categorized using the SCORPAN methodology, a set of best performing prediction methods and an analysis of the validation strategies used in similar studies. The results of the literature review are used for designing the artifact (prediction method). The second part of the study is the design and development of the artifact. During this research, we have followed the steps of Design Science Research Methodology (DSRM) and used the environmental covariates found in the literature review. We have tested three machine learning algorithms and Random Forest had the highest predictive accuracy for this data set (R-squared = 0.37, MSE = 1.76). The developed method serves as a guideline for soil organic carbon prediction in the Netherlands. However, in future research, this method should be evaluated with a larger data set to prevent overfitting.