Remote sensing of seagrasses in a patchy multi-species environment

Abstract

We tested the utility of IKONOS satellite imagery to map seagrass distribution and biomass in a 4.1 km² area around Chumbe Island, Zanzibar, Tanzania. Considered to be a challenging environment to map, this area is characterized by a diverse mix of inter- and subtidal habitat types. Our mapped distribution of seagrasses corresponded well to field data, although the total seagrass area was underestimated due to spectral confusion and misclassification of areas with sparse seagrass patches as sparse coral and algae-covered limestone rock. Seagrass biomass was also accurately estimated (r ² = 0.83), except in areas with Thalassodendron ciliatum (r ² = 0.57), as the stems of T. ciliatum change the relationship between light interception and biomass from that of other species in the area. We recommend the use of remote sensing over field-based methods for seagrass mapping because of the comprehensive coverage, high accuracy and ability to estimate biomass. The results obtained with IKONOS imagery in our complex study area are encouraging, and support the use of this data source for seagrass mapping in similar areas.

Acknowledgements

We thank the staff of Chumbe Island Coral Park for their cooperation and good company, and the Institute of Marine Sciences on Zanzibar for access to laboratory facilities. Funding for satellite imagery was provided by Ellsworth LeDrew, and sound advice was provided by Martin Gullström and Bengt Lundén. Constructive comments were provided by two anonymous reviewers.