engleski

The applicability of various seabed classification systems in mapping spatial distribution of the seagrass Posidonia oceanica (L.) Delile

The purpose of this research was to determine the numerical precision of various cost-effective benthic habitat mapping techniques and their suitability for mapping and monitoring of Posidonia oceanica meadows in the Croatian Adriatic. Ultra–high resolution aerial imagery, single-beam echo sounder (SBES) seabed classification system from Quester Tangent Co. (QTC), and surface based underwater videography were selected as affordable, non-destructive and simple to use systems for data acquisition. The ultra-high resolution digital imagery was capable of detecting P. oceanica meadows up to 4m depth with 94% accuracy, from 4m to 12.5m depth the accuracy dropped to app. 76%, and from 12.5 to 20m the software was only capable of distinguishing seabed biota from substrata, with 97% accuracy. The results of the QTC system revealed over 90% detection accuracy of Cymodocea nodosa covered seabed, excellent separation capabilities (>92%) of different sediment types, and reasonable accuracy for mapping underwater vegetation regardless of the bathymetric span. The system proved incapable of separating P. oceanica from dense macroalgae on the same type of substratum. Surface-based underwater videography demonstrated great potential for estimating P. oceanica cover in a sampled region using either a single human rater or a computer estimate. The consistency between two human scorers in evaluating P. oceanica bottom coverage was near perfect (>98%) and high between digital and human scorers (80%).

Habitat mapping; Remote sensing; Bottom coverage; Computer scoring; System integration; Posidonia oceanica; Adriatic Sea

nije evidentirano