A master thesis from Aalborg University
En konceptuel model til geografisk overvågning af jordbundssundhed
[A conceptual model for geographical monitoring of soil health]
Forfatter(e)
Semester
4. semester
Uddannelse
Udgivelsesår
2024
Afleveret
2024-10-20
Antal sider
160 pages
Abstract
This report addresses the need to monitor soil health to ensure food security in the face of a growing global population and climate change. The study utilises remote sensing data, including Sentinel-2 satellite imagery, CORINE land use data, climatic variables, and terrain information. PyCaret in Python is used to develop the machine learning models, enabling efficient model training and comparison of various algorithms. These predictive models are integrated into a broader conceptual framework designed for large-scale soil health monitoring. By combining multiple data sources, this framework ensures a comprehensive assessment of soil conditions across diverse regions, making it a valuable tool for evaluating soil health at a large geographical scale. A key component of this approach involves using pedotransfer functions to estimate important soil properties, such as hydraulic conductivity and plant-available water, which are not directly measured. These estimates are included in the multi-criteria analysis, ensuring a holistic evaluation that incorporates both direct and inferred data. The models are calibrated with soil samples to validate their accuracy, ensuring robust predictions across broad geographic regions. Various machine learning models are compared based on accuracy, using metrics such as RMSE, to minimize deviations in the subsequent multi-criteria analysis. This process integrates multiple factors, to make a comprehensive assessment of soil health and supporting large-scale monitoring efforts. Although machine learning proves effective for predicting soil conditions, the complexity of soil composition due to interacting factors remains a challenge. The study concludes that large parts of the Iberian Peninsula show poor soil health, while smaller areas, particularly in the east, exhibit acceptable to good conditions. This model and methodology thus offers a conceptual framework for future large-scale soil health monitoring.
Emneord
Kolofon: Denne side er en del af AAU Studenterprojekter — Aalborg Universitets studenterprojektportal. Her kan du finde og downloade offentligt tilgængelige kandidatspecialer og masterprojekter fra hele universitetet fra 2008 og frem. Studenterprojekter fra før 2008 kan findes i trykt form på Aalborg Universitetsbibliotek.
Har du spørgsmål til AAU Studenterprojekter eller Aalborg Universitets forskningsregistrering, formidling og analyse, er du altid velkommen til at kontakte VBN-teamet. Du kan også læse mere i AAU Studenterprojekter FAQ.