See Seagrass from Space: From Analysis-Ready PlanetScope Satellite Imagery to Nationwide Seagrass Maps for the Nationally Determined Contributions of Seychelles

Abstract

Seagrass meadows provide many valuable global ecosystem services including blue carbon sequestration, and habitat and nursery ground provisioning for commercial fish. Yet, the conservation, research and monitoring of these highly productive habitats are hampered by a notable lack of spatially-explicit knowledge in many parts of the world, including the Seychelles. As an archipelagic state, a sizeable portion of the Seychellois Nationally Determined Contributions and thus carbon stocks are “blue”, owing to the therein intertidal and subtidal seagrasses. Due to the subtidal seagrass living on the seabed, the dominant spaceborne remote sensing approach is optical. However, owing to the atmospheric conditions and persistent cloud cover over tropical regions, it is sometimes difficult to find a suitable, cloud-free and glint-free image. In order to maximise use of every relevant and useful pixel without discarding them due to a partial cloud cover, we employ a multitemporal image composition approach, which concurrently reduces some of the effects of sunglint and cloud shadows. However, a multitemporal approach on a national scale is computationally expensive and requires much resources. Here we present the Global Seagrass Watch coastal Ecosystem Accounting framework, which harnesses the powerful cloud computing Google Earth Engine (GEE) platform to process the satellite images across a large spatial and temporal scale. Furthermore, with the addition of the Planet & Norway’s International Climate and Forests Initiative (NICFI) onto GEE, anyone can have free access to the 4.77 m high spatial resolution PlanetScope image composites between 2015 to present. Although terrestrially-focused, the presence of shallow waters within the buffered coastline in these composites allows the exploitation of NICFI for coastal aquatic remote sensing, as well. Consequently, a potentially high-resolution, large-scale seagrass map could be produced with savings in time and cost. In this study, we implement our multitemporal composition approach on GEE onto the PlanetScope composite archive in order to map and estimate the area of the seagrass meadows of Seychelles on a national level. The results will contribute to the development of a national seagrass mapping and monitoring blueprint monitoring system, which is important for the assessment of national seagrass blue carbon stocks for the Nationally Determined Contributions for Seychelles, as well as beyond.