Vegetation Biomass of Sargassum Meadows in An Chan Coastal Waters, Phu Yen Province, Vietnam Derived from PlanetScope Image

Abstract

SW (Seaweed) is a valuable coastal resource for its use in food, cosmetics, medicine and other items. In this study, PS (PlanetScope) imagery was combined with field sampling to demonstrate the capability of mapping of SAV (Submerge Aquatic Vegetation) (including both SW and SG (Seagrass) beds) and biomass mapping of Sargassum meadows in An Chan coastal waters, Tuy An district, Phu Yen province, Vietnam. In term of SAV and Sargassum mapping, authors proposed an improved remote sensing technique based on Sagawa’s BRI (Bottom Reflectance Index) algorithm with attention to Tassan’s concept in discrimination of light attenuation coefficient Kd between shallow and deep waters. Authors’ results showed high accuracy in mapping of SAV and Sargassum distribution with overall accuracy and Kappa coefficient of 92.52% and 0.8957, respectively. The classified class of SW (i.e. Sargassum sp.) then was separated absolutely from other classes in SAV map for estimation of Sargassum biomass. The red and green spectral pre-processed BRI channels (i.e. BRI3 and BRI2) of PS were used to estimate the Sargassum biomass using a multiple 2nd order polynomial regression model with very high accuracy (R = 0.9707; RMSE =  109.21 g/m). The average total Sargassum biomass was 897.8 g/m with total Sargassum yield in whole region reaching a value of 449.57 tons in cover area of 50.32 ha of Sargassum meadows. This result opens the great potential of biomass and yield estimation of Sargassum and other SW meadows in coastal waters (including enough optically deep waters) by remote sensing techniques based on PS imagery.