Affiliation(s)
1. Graduate University of Science and TechnologyVietnam Academy of Science and Technology, Ha Noi 11355, Viet Nam
2. Center for Environmental Monitoring, Department of Natural Resources and Environment, Phu Yen Prov. 56000, Viet Nam
3. Institute of Oceanography, Vietnam Academy of Science and Technology, Khanh Hoa Prov. 57000, Viet Nam
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 (R2 = 0.9707; RMSE = ± 109.21 g/m2).
The average total Sargassum biomass
was 897.8 g/m2 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.
KEYWORDS
PS satellite image, Sargassum biomass, SW, BRI.
Cite this paper
References