Mapping Seagrass Ecosystems in Norway with the YellowScan Navigator LiDAR
Challenge
The Norwegian Institute for Water Research (NIVA) is the leading partner of SeaBee, Norway’s infrastructure for drone-based research, mapping, and monitoring in the coastal zone. As part of this initiative, NIVA aimed to study seagrass ecosystems, which capture and store large amounts of CO₂. The goal was to investigate how much carbon is stored and to better understand the effects of climate and human impacts on these key submerged coastal habitats.
Traditional monitoring methods lacked the resolution needed for shallow, dynamic coastal waters. To address this, NIVA set out to create a three-dimensional model of a submerged seagrass bed using topo-bathymetric LiDAR, in order to quantify seagrass habitat volume and estimate carbon stored in above-ground biomass (leaves).
As a research institute, NIVA also needed a solution that was accessible and easy to use, providing robust hardware and software that could be operated by scientists and technicians without specialized training in LiDAR.
Photos from field deployment of Navigator, provided by Robert Nøddebo Poulsen.
Solution
To achieve this, NIVA deployed the YellowScan Navigator, a UAV-mounted topo-bathymetric LiDAR system, paired with a Hexadrone Tundra 2 drone equipped with endurance rotor arms. The Navigator was complemented with CloudStation Colorization and Terrain Classification modules, enabling intuitive data processing and accurate separation of vegetation from seabed.
The system proved accessible and easy to use for the team of scientists and technicians, while YellowScan’s support provided timely feedback on instrument settings and troubleshooting during the campaign.
Screen captures of the point cloud at the study area.
Mission parameters
Survey size: 0.09 km² – 12.64 million points (50 m elevation); 0.05 km² – 12.65 million points (25 m elevation)
- Flights: 2 total
- Flight time: 24 min (50 m) | 22 min (25 m)
- Flight altitude & speed: 25 m and 50 m | 5 m/s
- Equipment:
- YellowScan Navigator + camera module
- Tundra 2 UAV (Hexadrone, France)
- CloudStation Colorization & Terrain Classification
- CloudCompare (for manual inspection/cleaning)
Screen capture of the point clouds at the study area.
Screen capture of the point clouds at the study area.
Results
The YellowScan Navigator delivered point densities of 70 pts/m² at 50 m AGL and 125 pts/m² at 25 m AGL, with horizontal and vertical accuracies between 0.01 and 0.015 m depending on flight height.
Using the Terrain Classification feature in the YellowScan CloudStation software, the raw point cloud was separated into land, water surface, water column noise, and sea floor. After manual cleaning, only seafloor points were retained and further refined with a Random Forest model, where point intensity was the most important variable, to distinguish vegetation from true seabed (sand and rock).
This workflow enabled the creation of high-resolution Digital Surface Models (DSMs) of the vegetation canopy and Digital Terrain Models (DTMs) of the seabed, both at 10 cm resolution. Subtracting the DTM from the DSM produced the first canopy height models of submerged seagrass at this level of detail, allowing NIVA to quantify habitat volume and estimate carbon capture in above-ground biomass.
Scientifically, the Navigator provided point cloud densities an order of magnitude higher than airborne LiDAR surveys, enabling unprecedented 3D modeling of submerged vegetation. The system proved accurate, reproducible, cost-effective, and intuitive to operate, while CloudStation’s classification and intensity correction tools significantly expedited processing. Compared to airborne LiDAR, the drone-borne Navigator allowed surveys that were not only cheaper and safer to deploy, but also more easily reproducible for future shallow-water habitat monitoring campaigns.
The YellowScan Navigator LiDAR system seems ideal for researchers and natural resource managers alike, with intuitive and rapid data processing of high-quality point cloud data. This system represents the next step forward for shallow water habitat mapping, facilitating cost-effective, reproducible, and high-resolution results relative to traditional surveying methods.
Canopy height model imposed over an RGB drone image of the study site, including the surveyed eelgrass bed.
