University of Twente Student Theses

Login
As of Friday, 8 August 2025, the current Student Theses repository is no longer available for thesis uploads. A new Student Theses repository will be available starting Friday, 15 August 2025.

Estimation of Micronutrients in Maize yield in Malawi using Machine Learning and spatially explicit environmental data

Bian, Xinran (2025) Estimation of Micronutrients in Maize yield in Malawi using Machine Learning and spatially explicit environmental data.

[img] PDF
2MB
Abstract:Micronutrient deficiencies in staple crops contribute to “hidden hunger” across sub-Saharan Africa, yet field measurements of grain nutrient content remain spatially sparse and costly. This study develops a framework for predicting four key micronutrients, calcium (Ca), iron (Fe), zinc (Zn) and selenium (Se), in Malawian maize grain by combining one-time ground observations with 61 terrain, climate, soil, remote-sensing variables, and Global Agro-Ecological Zones. An XGBoost algorithm was trained for each micronutrient using spatial cross-validation approach to address the potential spatial autocorrelation bias. Model skill varied: test R² reached 0.15 for Ca, 0.14 for Zn, 0.08 for Fe and was negligible for Se; RMSEs were 18.9 mg kg⁻¹ (Ca), 25.1 mg kg⁻¹ (Fe), 3.7 mg kg⁻¹ (Zn) and 0.03 mg kg⁻¹ (Se). Elevation, precipitation, top-soil pH and organic carbon strongly influenced Ca, Fe and Zn predictions, while radar-derived canopy metrics dominated Se. The spatial distribution of the maize nutrient revealed a south-to-north decrease in micronutrient levels, aligning with the known gradients in soil pH, organic matter and climate. By exploiting low-cost geodata, the framework can guide site-specific fertiliser use, deployment of biofortified varieties and nutrition programmes, helping Malawian authorities combat hidden hunger and progress toward Sustainable Development Goal.
Item Type:Essay (Master)
Faculty:ITC: Faculty of Geo-information Science and Earth Observation
Subject:43 environmental science
Programme:Geo-informatics MSc (60031)
Link to this item:https://purl.utwente.nl/essays/107286
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page