Mapping the Intertidal Zone Using Multispectral Imagery Retrieved by Unmanned Aerial Vehicles in the Northwest Coast of Portugal

In-situ conventional surveys for biodiversity assessment are time and resources consuming, calling for the development of innovative, reliable yet expeditious, tools to map the intertidal using remotely sensed images. The present work is being developed in the scope of the project SWUAV. The basis of SWUAV approach is the discrimination of four intertidal components: seaweed (SE); sand (SA); mussels and rock (MR); rock, barnacles and limpets (RBL) by semi-automatic classification techniques applied to multispectral imagery data in QGIS software. UAVs equipped with RGB and multispectral cameras, performed flights at 20 m height, yielding pixel sizes of 0.8 and 1.6 cm, respectively. Flights were done in February and May 2019 to assess temporal variation of those four components as a part of a pilot study in a rocky shore located in the NW coast of Portugal. All Images were orthorectified, mosaicked and precisely georeferenced using control points surveyed with a high accuracy GPS receiver, allowing data of different sensors and different periods to be overlaid in a Geographic Information System. Supervised classification based on NDVI reflected the expected seasonal growth of seaweed beds and gave insights on variation of relative percentage of main benthic invertebrates. From February to May, decreases of 18% and 27% were observed in areas of SA and RBL, while for MR and SE, areas increased by 54% and 31%, respectively. In February the overall accuracy of the supervised classification was 71% and kappa 0.6, while in May it reached 76.6 % and had a kappa value of 0.7. The information on the calculated area for seaweed beds is currently being crossed with data on species spectral discrimination and biomass/NDVI variation, so as to allow for the remote assessment of the seaweed biomass. Future work in the scope of SWUAV will combine this information with hyperspectral data and aims to develop a QGIS plugin to remotely assess and map intertidal seaweed biomass, thereby improving the knowledge of the local ecosystems, and contributing for efficient and sustainable management of these coastal habitats.