Dragonfly team member Fabian Döweler, PhD, is mapping regenerating forests in Aotearoa with the help of the Ministry of Business, Innovation and Employment’s Catalyst Fund. The goal is to be able to monitor change in the forests in response to management, such as pest control, or after destructive events, such as fire.
Building on work done in collaboration with Vibrant Planet, Fabian is working with NASA on a technology that uses Artificial Intelligence (AI) to produce accurate models of tree canopies, combining LiDAR (Light Detection And Ranging) with aerial imagery.
LiDAR data is currently the gold standard for building forest structure models —but it’s time-consuming and expensive data to gather at scale.
“Imagine LiDAR as a high-tech echolocation system, but instead of sound, it uses light. Picture sending out millions of tiny laser pulses, like thousands of ping-pong balls bouncing off the forest canopy. Some of these balls hit the treetops, while others penetrate through branches and leaves, eventually reaching the ground,” Fabian says.
“By measuring how long it takes each ball to bounce back, we can create a three-dimensional map of the forest, revealing not just the height of the trees, but also the density and structure of the canopy—like peeling back the layers of a complex, living puzzle.”
Monitoring forest carbon
Fabian says the need for accurate, regularly updated information on the Earth’s forests is crucial for the management of carbon stocks and ultimately carbon emission reduction.
There are both large- and smaller-scale applications of his work. Some examples include modelling carbon sequestration across large areas during normal growth phases, but also monitoring growth after a destructive event, or after the control of pests like possums.
NASA Earth observations enable monitoring of New Zealand’s forests
Dragonfly Data Science’s capability to capture fine detail about the forest structure complements NASA’s ability to do large-scale forest monitoring through the agency’s spaceborne Earth observation programme.
Fabian is currently working on two use cases for his tree canopy modelling alongside NASA researchers, including Dr Taejin Park, Research Scientist, with NASA Earth eXchange (NEX)
The NASA team has tested a forest-monitoring system over New Zealand which they’ve initially developed for Mexico under support from NASA CMS (Carbon Monitoring System) programme.
The system brings together NASA’s unique Earth observation assets including spaceborne LiDAR GEDI (Global Ecosystem Dynamics Investigation) and multispectral Landsat imageries.
During this feasibility phase, the project team successfully implemented the monitoring system over the Coromandel region in New Zealand for the Department of Conservation, modelling forests in this North Island region while the agency works to remove pests such as possums from the area.
The team expects that with the removal of possums, forest health will markedly improve (this pest is particularly destructive to trees). Fabian’s mapping work will be one of the ways the team can monitor the benefits of pest control.
The second application is on the Caldor fire zone in California, which was significantly damaged during the 2021 wildfires. There are no pre-fire LiDAR data, so the team is using NEX satellite-based historical maps in combination with airborne imagery to create the canopy models.
This Catalyst project is expecting first results in the coming months.