From In Situ to satellite observations of pelagic Sargassum distribution and aggregation in the Tropical North Atlantic Ocean

2019
The present study reports on observationscarried out in the Tropical North Atlantic in summer and autumn 2017, documenting Sargassumaggregations using both ship-deck observationsand satellite sensor observationsat three resolutions (MSI-10 m, OLCI-300 m, VIIRS-750 m and MODIS-1km). Both datasets reported that in summer, Sargassumaggregations were mainly observedoff Brazil and near the Caribbean Islands, while they accumulated near the African coast in autumn. Based on in situ observations, we propose a five-class typology allowing standardisation of the description of in situ Sargassum raftshapes and sizes. The most commonly observed Sargassum rafttype was windrows, but large raftscomposed of a quasi-circular patch hundreds of meters wide were also observed. Satellite imageryshowed that these raftsformed larger Sargassumaggregations over a wide range of scales, with smaller aggregations (of tens of m2 area) nested within larger ones (of hundreds of km2). Match-ups between different satellite sensors and in situ observationswere limited for this dataset, mainly because of high cloud coverduring the periods of observation. Nevertheless, comparisons between the two datasets showed that satellite sensors successfully detected Sargassumabundance and aggregation patternsconsistent with in situ observations. MODIS and VIIRS sensors were better suited to describing the Sargassumaggregation distribution and dynamics at Atlantic scale, while the new sensors, OLCI and MSI, proved their ability to detect Sargassumaggregations and to describe their (sub-) mesoscale nested structure. The high variability in raftshape, size, thickness, depth and biomass density observedin situmeans that caution is called for when using satellite maps of Sargassumdistribution and biomass estimation. Improvements would require additional in situand airborne observationsor very high-resolution satellite imagery.
    • Correction
    • Source
    • Cite
    • Save
    57
    References
    23
    Citations
    NaN
    KQI
    []
    Baidu
    map