WORKSHOP - Technologies for Assisting Forest Ecosystem Restoration
On February 19th-21st, 2025, a small group of scientists, engineers and conservationists gathered at the Doi Suthep Nature Education Centre to address an important question: how can technologies help with upscaling tropical forest restoration? This question has become particularly timely, since the launch of RESULT Asia-Pacific , FAO’s plan to reforest 100 million hectares by 2030.
The workshop, entitled "Technologies for Assisting Forest Ecosystem Restoration", brought together experts from Chiang Mai University (CMU), RMIT Vietnam, the University of Glasgow Singapore (UGS), King Mongkut’s University of Technology Thonburi (KMUTT), University of Oulu and Bioversity International , among others.
The event was part of the FIREfly 2.0 project, a collaborative initiative, focused on enhancing forest-fire monitoring and biodiversity conservation, using both ground and UAV sensors and advanced communication systems. Funding was provided by the APNIC Foundation through the ISIF Asia Grant for Internet and Environment.
The workshop was conceived as an opportunity to identify synergies among researchers, who are working on similar technologies, and to share ideas and methods to maximize impact.
Fighting Fire with Technology
Wildfires are one of the greatest obstacles to successful restoration, threatening to erase decades of hard work in a few hours. On the first day of the workshop, several contributors presented cutting-edge solutions for early fire detection, prediction and suppression.
Dr. Dinesh Bhatia and Dr. Henrik Hesse (FIREfly Project, RMIT & UGS) introduced a low-cost sensor network for detecting forest fires. Their UAV-integrated system collects data from smoke detectors, humidity sensors and thermometers, relaying information via WiFi routers mounted on drones. A prototype has demonstrated the system’s ability to act as a fire alarm, which would allow firefighters to intervene before fires spread uncontrollably.
A critical issue, raised by several speakers, was flight time limitations of UAVs and the short distance over which signals can currently be transmitted. Dr. Kampol Woradit (CMU Engineering) is working on expanding WiFi range to 100 meters and LoRa to 200 meters, reducing dependence on drones for real-time communication.
Meanwhile, Naveen Babu Kanda presented machine-learning models that can predict high-risk fire areas using climate, topography and human activity data and thus help to plan the placement of fire detection and control measures.
Seeing the Forest for the Trees: New Tools for Monitoring Restoration
Once forests are restored, monitoring their growth, biodiversity and carbon storage is essential. Several teams presented remote sensing and deep learning approaches for tracking forest recovery.
Maxime Réjou-Méchain (IRD, France) demonstrated how hyperspectral imaging, LiDAR and deep learning models can map forest types, detect tree mortality, and track canopy growth. His team’s work in Khao Yai National Park has correlated LiDAR data with biomass estimates, providing highly accurate assessments of forest recovery.
At a finer scale, Dr. Watit Khokthong (CMU Environmental Science) and his team are using UAV-based LiDAR, combined with terrestrial LiDAR scans, to estimate leaf area index (LAI)—a key indicator of carbon sequestration. His findings are being integrated into Thailand’s carbon-credit verification system (TVERS).
Worayut Takaew presented his PhD plan to use drones to identify tree species from above, which could transform seed collection for propagating planting stock of indigenous tree species. By applying machine learning to high-resolution spatial and spectral data, his project aims to develop a rapid, scalable method for monitoring tree diversity and restoration success.
Roger Norum (University of Oulu) introduced the FORbEST project, a European Union-funded initiative with 18 partners, including CMU. The project will explore forest management best practices, biodiversity monitoring and carbon tracking, using tools like eDNA, LiDAR, mobile apps and eco-acoustic recorders.
Manuela Hirschmugl (Joanneum Research) showed how remote sensing tools, including ALS, GEDI and Sentinel-2 data, can be used to assess restoration indicators such as forest structure, connectivity, deadwood and disturbance levels. She recommended tracking FHD (foliage height diversity) and sudden changes in NIR reflectance—which signal forest loss, regeneration extent and disturbance causes.
Stefan Prost (University of Oulu) introduced eDNA sampling techniques for biodiversity monitoring. By analysing trace DNA left in soil and on plants, his team plan to identify species diversity in restored forests without direct observation. However, he stressed the need to expand DNA reference databases, as many tropical species have yet to be DNA sequenced.
Drones—the Future of Forest Restoration?
Perhaps the most ambitious concept at the workshop was the use of drones to replace tree planting with aerial seeding.
Dr. Annop Ruangwiset (KMUTT) contributed a presentation describing a drone-mounted seed hopper designed to distribute seeds. His team is exploring the use of multiple seed-dispersal mechanisms, including gravity-drop systems, fluted rollers and air-blown dispersal.
However, Dr. Khuanpirom Nairuangsri (FORRU-CMU) highlighted the need to consider the biology of the seeds themselves—not just mechanisms to disperse them. Drone-dropped seeds face a multitude of seed predators and those that escape predation have low germination rates. Rodents, ants and harsh environmental conditions destroy many seeds and young seedlings, long before they can grow into trees. Her field trials found that while one species achieved germination rates above 50% (Adenanthera microsperma), many failed to germinate entirely. Seed balls, made of biochar, may deter predators and desiccation, but they also appear to inhibit germination by reducing penetration of light, water and/or oxygen into the seed embryo.
Day 2 of the workshop consisted of field demonstrations at FORRU-CMU's restoration plots near Ban Mae Sa. Participants observed practical applications of terrestrial LiDAR and the use of drones for monitoring, communication and seed dropping.
Following presentation of ground-based methods of forest restoration and data collection, Watit Khokthong and Worayut Takaew (CMU) showcased how Terrestrial LiDAR can generate high-resolution 3D models of forest structure, within which accurate measurements of trees height, density and biomass can be made.
The CMU Engineering team demonstrated their data-mule drone, showcasing how drones can collect environmental data from ground smoke detectors. This technology reduces the need for manual data retrieval and ensures continuous monitoring. Worayut Takaew flew an RTK multispectral drone on a mission to collect images in 6 wavebands, which he intends to use to identify trees species from their spectral signatures. The field day concluded with the successful flight of a seed-dropping drone prototype.
Day 3 of the workshop, back in the DSNC, explored various software applications, which enable detailed spatial analysis and 3D reconstruction of forest sites from imagery: Pix4D, Cloud Compare, Agisoft Metashape and R packages. Such tools can help to devise restoration strategies and track subsequent progress. Applications, to assist with species selection and restoration planning were also introduced. Riina Jalonen (Bioversity International) explained how the D4R (Diversity for Restoration) app helps practitioners to select tree species that are well-matched with both site conditions and restoration objectives. The system is currently being modified to include northern Thailand . With a radically new approach, Kornprom Pikulkaew (CMU Mathematics) proposed developing a system that would use AI to derive optimum forest-restoration strategies simply by scanning photographs of restoration sites.
The Road Ahead: An Integrated Approach to Tech-Assisted Restoration
What became apparent during the workshop is the need to integrate various technologies. Collaboration was also seen as essential to meet the technological challenges of restoring tropical forest ecosystems on the vast scales currently envisaged by ambitious region-wide initiatives such as RESULT .
Discussions during the workshop explored synergies between remote sensing, field monitoring and software analyses, thus identifying opportunities for sharing of data, equipment and other resources, to improve restoration research outcomes. If these innovative projects align their efforts, they might just revolutionize the way tropical forests are restored.
Stephen Elliott 22/2/25
