This master’s thesis investigates the prediction of vegetation dynamics in the Troodos Mountains using time series analysis of the Normalized Difference Vegetation Index (NDVI). NDVI is a widely used remote sensing metric that quantifies vegetation greenness and health, making it a valuable tool for monitoring ecological changes, land cover, and environmental disturbances. The study is situated within the context of the Troodos Mountains, a region of ecological and climatic significance in Cyprus, and aims to leverage historical NDVI data to understand and forecast vegetation trends over time.
Comprehensive NDVI Time Series Analysis: The thesis compiles and analyzes NDVI data collected over multiple years for the Troodos Mountains. By examining temporal patterns, the study identifies both seasonal and interannual variations in vegetation cover, providing a detailed picture of ecological dynamics in the region.
Application of Predictive Modeling: Advanced statistical and machine learning techniques are employed to model and predict future vegetation dynamics based on historical NDVI time series. This includes trend analysis, anomaly detection, and the use of predictive algorithms such as long short-term memory (LSTM) networks, which are particularly suited for sequential data and have shown promise in ecological forecasting.
Assessment of Environmental Drivers: The research explores the relationship between NDVI trends and environmental variables such as climate, land use, and disturbance events (e.g., wildfires, droughts). By correlating NDVI fluctuations with these factors, the thesis enhances understanding of the drivers behind vegetation change in Mediterranean mountain ecosystems.
Regional Focus and Methodological Rigor: Focusing on the Troodos Mountains, the study contributes region-specific insights while employing robust data preprocessing, validation, and statistical testing to ensure the reliability of its findings.
The findings of this thesis have significant implications for environmental monitoring, land management, and climate adaptation strategies in Cyprus and similar Mediterranean regions. By demonstrating the utility of NDVI time series for detecting and forecasting vegetation changes, the research provides a methodological framework that can be adapted for other regions facing ecological pressures.
The predictive models developed in the study can inform policymakers and land managers about areas at risk of degradation or in need of conservation intervention. Furthermore, the integration of remote sensing data with advanced analytics supports the transition toward data-driven environmental decision-making. The thesis also contributes to the growing body of literature on the application of machine learning in ecological forecasting, highlighting both the opportunities and challenges of using satellite-derived indices for long-term environmental assessment.
In summary, this work advances the understanding of vegetation dynamics in the Troodos Mountains, showcases the power of NDVI time series analysis, and provides actionable insights for sustainable land and ecosystem management.