Perspectives in remote sensing for forests

Introduction

Forests are complex, dynamic systems that require advanced observation techniques to understand their structure, function, and changes over time. They play a vital role in regulating the global climate, supporting biodiversity, and sustaining the livelihoods of millions of people around the world. However, increasing pressures from deforestation, climate change, and land-use conversion demand new ways to monitor and model these ecosystems at multiple scales.

State of knowledge & Research topics

Recent advances in remote sensing and artificial intelligence have significantly altered the methods used to monitor and comprehend forests. The increasing availability of time series from satellite data and high-resolution aerial imagery, together with LiDAR, enables detailed analyses of forest structure and dynamics at varying scales, i.e. from stand to individual tree level. By merging information from various sensors, it is feasible to detect subtle changes in canopy height, crown structure, and spectral characteristics associated with tree species and health. This transition from stand-level to tree-level analysis marks a paradigm shift in forest monitoring, enabling more precise assessments of tree health, growth, and mortality.

Deep learning methods have emerged as powerful tools for automating these analyses, supporting applications from tree detection and crown delineation to species classification, biomass estimation, and the assessment of tree health and stress. These approaches are increasingly employed to detect and characterize forest disturbances such as wildfires, pest outbreaks, drought impacts, and windthrows, contributing to the development of early-warning and resilience assessment systems.

A central challenge remains the harmonization and fusion of heterogeneous data sources, as well as the scalability and interpretability of machine learning models across different forest types and ecological regions. Addressing these issues requires close collaboration between remote sensing scientists, ecologists, and forest managers. Our SIG will therefore promote interdisciplinary research and methodological innovation focused on reproducible workflows, open and harmonized datasets, and the transfer of cutting-edge analytical methods into operational forest monitoring systems. By advancing tree-level observation and analysis, the group aims to bridge the gap between technological development and practical implementation, supporting both precision forestry and evidence-based forest policy. The engaged strong collaboration between remote sensing researchers and forestry practitioners fosters the correct and intended use of remote sensing-based forest products.

Scope and objectives of the SIG

Building on this foundation, our Special Interest Group focuses on advancing and operationalizing remote sensing and artificial intelligence methods for forest monitoring, management and research. Its main areas of interest include:

Multi-scale forest observation using multispectral, hyperspectral and LiDAR data.

Tree-level analysis for mapping forest structure, species composition, health, and biomass.

Application of deep learning and machine learning for classification, disturbance detection, and resilience assessment.

Multi-sensor data fusion, multi-temporal analysis, and harmonization of heterogeneous data sources.

Development of reproducible, open, and interoperable workflows for operational forest monitoring.

Explainable and trustworthy AI to enhance interpretability and user confidence in forest information products.

Benchmarking, validation, and creation of harmonized datasets for comparative and cross-regional studies.

Translation of research outcomes into operational tools for forest management, conservation, and policy support.

To strengthen collaboration and knowledge exchange, the SIG will work closely with other international societies and working groups (e.g., within the ISPRS and IUFRO communities), promote joint scientific publications such as special issues in peer-reviewed journals, and support training and outreach activities that connect researchers and practitioners across disciplines.

Workshops & SIG activities

We plan to organize sessions or workshops on a yearly basis both in-person and online. These workshops will be organized back-to-back with other events, such as the EARSeL symposia, IUFRO meetings or another forest and/or remote sensing related event.

Previous workshops

Prague, Czech republic (26th May, 2025)
The vision of the 4th Workshop on Forestry was to gather the current state-of-the-art in RS to support forestry-related policies and to jointly identify remaining gaps and the way forward. Remote sensing based forest monitoring in support of current and emerging policies. The workshop was organized by Manuela Hirschmugl and Janik Deutscher and it was dedicated to applications of remote sensing for the support of current and emerging European policies. The workshop was a great success with 22 presenters and around 100 participants from all around the world and an intriguing keynote by Annemarie Bastrup-Birk, University of Copenhagen.

Kraków, Poland (15-16 September, 2016)
3rd EARSeL SIG Forestry workshop: Breaking dimensions and resolutions of forest remote sensing data

Warsaw, Poland (17-18 June, 2014)
2nd EARSel SIG Forestry workshop: Remote Sensing for forestry applications

Prague, Czech Republic (2-3 June, 2011)
1st EARSeL SIG Forestry workshop: Operational remote sensing in forest management – GRSS | IEEE | Geoscience & Remote Sensing Society

General platforms for forestry data or monitoring platforms

If you know datasets, websites, etc. that provide open data, please email us and we will add them to our list.

Name	Description & usage

deadtrees.earth	An open database for accessing, contributing, analyzing, and visualizing remote sensing-based tree mortality data.
TreeAI	Data call for tree species annotations and aerial RGB or RGBI imagery.
FOR-species20K dataset	Data for benchmarking tree species classification from proximally-sensed laser scanning data.
European Forest Disturbance Atlas	This repository holds maps of annual forest disturbances across 38 European countries derived from Landsat satellite data.