Detecting Sargassum Bloom Directly From Satellite Top-of-Atmosphere Reflectance With High-Resolution Images.

Massive floating Sargassum blooms have occurred frequently in many parts of the global ocean, and satellite remote sensing provides an effective way to monitor their spatiotemporal variation. Coarse-resolution satellite data, however, often suffer from a data gap in nearshore waters and detection uncertainty of Sargassum amount, especially for smaller patches. These limitations may be ameliorated by high-resolution satellite data, yet such great potential is hindered by the lack of reliable and easy-to-implement methods to detect Sargassum slicks. Here, combining the visible and near-infrared (NIR) topof-atmosphere reflectance ( R-TOA) data with the random forest (RF) model, a new method, namely the RTOA-RF model, was designed to automatically detect Sargassum from high-resolution satellite imagery with four wavebands. Specifically, this model was successfully applied to various satellite sensors, including Gaofen1-wide field view multispectral camera (GF1-WFV; 16 m), GF2 multispectral scanner (GF2-MSS; 4 m), GF6-WFV (16 m), Huanjing1A/B charge-coupled device (HJ1A/B-CCD; 30 m), Haiyang1C/D coastal zone imager (HY1C/D-CZI; 50 m). Comparisons with visual inspection and cross-index indicated that all achieved satisfactory performance for detecting Sargassum slicks, with overall accuracy (OA) and Kappa values greater than 96% and 88%, respectively. The sensitivity analysis of the model as an example of different sensors suggested that the RTOA-RF model can effectively identify Sargassum features under complex ocean background, cloud cover, and sunglint, even under conditions of surface wave glitter and weak Sargassum feature with low false positive rate (<9.60%). The findings here not only pave the way for operational monitoring of fine-scale Sargassum, but importantly provide thought for developing detection method of macroalgae blooms in global water.