UAV visual imagery-based evaluation of blue carbon as seagrass beds on a tidal flat scale

Abstract

Seagrass and seaweed beds (SSBs) have a high carbon sequestration function (blue carbon) in shallow coastal waters. Unmanned aerial vehicles (UAVs) are a highly useful tool for monitoring SSBs because of their ease of use and ability to acquire high-resolution photographs. In many previous studies using UAV, surveys of SSBs have been based on area alone, but it is insufficient to properly assess the habitat and carbon fixation of SSBs.

In this study, we estimated above-ground biomass and carbon of eelgrass in shallow coastal waters by combining aerial photography of visible images, quadrat surveys, and sampling of eelgrass. The analysis area was a tidal flat on an island located in the Seto Inland Sea in western Japan. Aerial photography was conducted by UAV to acquire high-resolution RGB visual images of the area. The quadrat survey and sampling were used to develop regression formulas for estimating biomass and carbon of eelgrass. The former was conducted to investigate the relationship between the coverage and Leaf Area Index (LAI), and the latter was conducted to investigate the relationship between leaf area and biomass, carbon of eelgrass. Those showed clear relationship between coverage and LAI (R² = 0.97) and between leaf area and biomass, carbon (biomass: R² = 0.98, carbon: R² = 0.98).

To identify eelgrass beds, the maximum likelihood classification was adapted. After calculating the coverage from the distribution, biomass and carbon were estimated by adapting regression formulas developed by quadrat survey and sampling.

The proposed method can be easily adapted from visible images taken by UAVs and robust to the effects of water, which provides high adaptability regarding the estimation for biomass and carbon of eelgrass on the tidal flat.