Drone-Based Spatio-Temporal Assessment of a Seagras Meadow: Insights into Anthropogenic Pressure

Abstract

Zostera marina L. is a flowering plant of great ecological interest as a breeding, nursery, and feeding place for many species. However, its spatial location implies strong competition with human uses (boating, fishing, etc.). Regular monitoring at a very high spatial and temporal resolution by a drone has been initiated to study the spatio-temporal and ecological dynamics of the seagrass meadow. Three drone campaigns per year were carried out in 2021 and 2022, totaling six spatial models. A pixel-oriented classification was performed to determine the overall envelope and to analyze the fragmentation of the meadow, which is likely caused by anchorage. A yearly loss of 465.18 m2 was measured (envelope area) and a difference of 12.15 m2 was observed between 2021 and 2022 (fragmented envelope area).

Keywords

• seagrass
• drone
• machine learning