An approach for quantification of submerged seagrass biomass in shallow turbid coastal waters

Abstract

In this paper, we introduced a new approach for absolute seagrass biomass determination using remote sensing technique. Two-step data processing of satellite remote sensing data was performed, namely: (i) detection and mapping of seagrass occurrence; and (ii) determination and mapping of seagrass biomass. The newly launched Landsat-8 Operational Land Imager band 2, 3 and 4 were used and applied in Merambong Shoals located in Straits of Johore, Malaysia that have various density of seagrass coverage in the region. The results indicated good agreement with in-situ verifications. The seagrass occurrence areas were classified at 90 percent overall classification accuracy (khat statistics=0.87) whilst the seagrass aboveground biomass was succesfully modelled agianst the in-situ observations (R2=0.9, p<0.0001, n=1200). It is then extended to fully map the above seagrass biomass for the large coverage with good accuracy (RMSE+1.75g/m2).