|Towards a better understanding of grass bed dynamics using remote sensing at high spatial and temporal resolutions |
Auteur(s): Marion Menu, Guillaume Papuga, Frédéric Andrieu, Guilhem Debarros, Xavier Fortuny, Samuel Alleaume, Pitard E.
Ref HAL: hal-03125706_v1
Exporter : BibTex | endNote
Wetlands conservation and resilience capacities are key issues in many places over the globe. Understanding these issues will benefit from a precise knowledge of seagrass species occupancy and coverage over time and over space. Such information can be obtained from remote sensing images and their classification thanks to a vegetation index, to be used in a complementary manner to field work inventories. Sentinel-2 data, which are available with a frequent revisit time (<5 days) and a high spatial resolution (10m pixel size) can be used to map grassbeds at the surface or slightly below the surface of permanent lagoons, hence enabling the characterization of its seasonal dynamics, which was not possible with previous remote-sensing tools. We have proved the feasibility of such a method in the natural reserve of the Bagnas (Herault, France) where Stuckenia pectinata coverage can be tracked over a full year thanks to Sentinel-2 images and field work. Inter-annual dynamics (seasonal growth and senescence) can be mapped over time with 10m resolution and will be extended to pluriannual studies thanks to the long-term objective of the Sentinel-2 mission. This opens the way to a concerted management of natural reserves based on data analysis and field knowledge, a better understanding of seagrass coverage with fluctuating environmental conditions, and predictive mechanistic and/or stochastic models of future qualitative trends.