University of Twente Student Theses
A Machine Learning Approach for Estimating Gross Primary Productivity Using Sentinel-2 Data
Karisma, K. (2024) A Machine Learning Approach for Estimating Gross Primary Productivity Using Sentinel-2 Data.
|
PDF
22MB |
| Abstract: | Predicting Gross Primary Productivity (GPP) across diverse ecosystems is essential for understanding the global carbon cycle and managing environmental resources effectively. This study evaluates the effectiveness of three different models, namely SARIMAX, XGBoost, and LSTM in estimating GPP using a combination of in-situ measurements and remote sensing data across various European ecosystems. The research consist of two main stages: the development of site-specific models to understand individual site characteristics and the creation of a unified model capable of generalizing predictions across different ecosystems without further site-specific adjustments. Our findings indicate that XGBoost consistently outperformed other models, showing superior prediction accuracy and robustness, particularly when generalized across multiple sites. SARIMAX and LSTM models also demonstrated useful capabilities, though with some limitations in specific contexts such as catastrophic forgetting in LSTM and poor performance in peak GPP predictions by SARIMAX. The inclusion of specific remote sensing indices, like the modified normalized difference vegetation index (MNDVI) and the enhanced vegetation index (EVI), significantly improved model performance across varied ecosystems. This study underscores the potential of integrating machine learning techniques with traditional ecological modeling approaches to enhance the prediction of GPP, which can significantly contribute to ecological management and climate change mitigation strategies. Future work should focus on refining these models’ ability to handle diverse data sets and improve their predictive reliability across global ecosystems. |
| Item Type: | Essay (Master) |
| Faculty: | EEMCS: Electrical Engineering, Mathematics and Computer Science |
| Subject: | 31 mathematics, 43 environmental science, 54 computer science |
| Programme: | Applied Mathematics MSc (60348) |
| Link to this item: | https://purl.utwente.nl/essays/102166 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page