Remote sensing characterization of benthic habitats and submerged vegetation biomass in Los Roques Archipelago National Park, Venezuela

Abstract

Remote sensing is a useful tool for characterizing submerged aquatic vegetation and other benthic habitats in shallow water areas with clear water transparency. In the present study, the visible bands of the Thematic Mapper (TM) sensor aboard Landsat 7 satellite were used in a supervised classification of benthic habitats and for the assessment of submerged vegetation biomass in Los Roques Archipelago National Park, Venezuela. Initially, the TM visible bands were log‐transformed and linearly combined to reduce the depth‐dependent variance in the bottom reflectance signal. The supervised classification had an overall accuracy of 74%. Eight bottom types could be spectrally separated: sand, dispersed communities over sand (shallow and deep), dense seagrass, dispersed seagrass meadows over sand, reef communities, mixed vegetation over muddy bottom, and lagoons. Regression analyses were performed between the depth‐invariant band combinations and field samples of vegetation biomass. The regression using the TM band 2 and 3 combination accounted for 64% of the variability of submerged vegetation biomass. According to these estimates, seagrass meadows with biomass between 64–96 g m−2 and 96–128 g m−2 predominate in the Los Roques Archipelago National Park.