Mapping forest felling activities in Latvia from Sentinel-2 satellite imagery using machine learning

Abstract

This thesis develops and evaluates a Random Forest (RF) machine learning classification framework for detecting clear-cut forest felling events using Sentinel-2 multispectral satellite data. While RF classification is widely applied in forest monitoring, its performance for operational clear-cut detection in Latvian semi-boreal forests remains insufficiently studied. This study aims to assess the applicability and robustness of RF model across multiple forest districts using multi-temporal satellite data.
Model training was conducted using four study areas within Latvian State Forests. For each area, cloud-masked Sentinel-2 imagery representing pre-felling (2018) and post-felling (2024) conditions was selected, resulting in multi-temporal datasets. Median spectral reflectance features were derived based on historical felling records.
Model evaluation included two approaches: internal validation within the training areas (self-testing), and external validation using a fifth forest district not included in training. Performance was assessed using overall accuracy, user accuracy (precision), producer accuracy (recall), and F1 score.
The results demonstrate that the RF model achieved over 70% user and producer accuracies in the unseen test area, with overall accuracy exceeding 95%. Analysis of feature importance showed that visible and shortwave infrared spectral bands contributed most significantly to distinguishing felled and non-felled areas. The inclusion of vegetation indices improved classification accuracy by 1–4%. Misclassifications were primarily related to sanitary fellings and cloud-contaminated pixels.
The findings indicate that Random Forest classification applied to Sentinel-2 data is capable of detecting clear-cut forest felling events in Latvian forests. The study highlights the importance of spectral band selection, vegetation indices, and robust validation strategies, contributing to improved forest monitoring capabilities in semi-boreal environments.

Popular Abstract

Forests are constantly changing due to natural processes and human activities such as logging, making monitoring these changes essential for sustainable forest management. To complement existing datasets and improve the scalability of forest monitoring systems, there is growing interest in using remote sensing and machine learning (ML) to detect and map clear-cut areas directly from satellite imagery. This study investigates the use of satellite imagery and machine learning for efficient detection of clear-cut forest areas.
Using freely available Sentinel-2 satellite images, a Random Forest machine learning method is applied to identify areas where forests have been cut. The approach focuses on several forest regions in Latvia and compares satellite images from before and after logging activities. By analyzing differences in how the forest reflects light, the model can distinguish between intact and harvested forest areas.
The results demonstrate that the Random Forest classification applied to Sentinel-2 data is capable of detecting clear-cut forest felling events in Latvian forests, reaching over 95% overall accuracy and more than 70% accuracy in identifying both felled and non-felled areas. The study also found that the most important information came from the visible and shortwave infrared satellite data bands, which are particularly effective for detecting forest changes. Adding vegetation indices, mathematical combinations of spectral data, further improves accuracy.
Classification errors were primarily associated with cloud contamination and other forest management practices, such as sanitary fellings, that exhibit spectral characteristics similar to clear-cut areas. Despite these challenges, the study demonstrates that satellite-based monitoring combined with machine learning is a powerful tool for tracking forest changes. This approach can support forest management, environmental monitoring, and policy-making by providing timely and accurate information about forest changes.

• author: Krumins, Gintars
• supervisor: Lars Eklundh
• organization: Dept of Physical Geography and Ecosystem Science
• course: GISM01 20261
• year: 2026
• type: H2 - Master's Degree (Two Years)
• subject: Earth and Environmental Sciences
• keywords: Latvia, Sentinel-2, Random Forest, remote sensing, GIS, forest felling detection, vegetation indices, spectral signature analysis.
• publication/series: Master Thesis in Geographical Information Science
• report number: 207
• language: English
• id: 9225256
• date added to LUP: 2026-04-13 11:05:00
• date last changed: 2026-04-13 11:05:00