Résumé
Satellite-linked buoys used by tropical tuna purse-seine vessels on drifting fish aggregating devices (DFADs) provide a continuous stream of information on both the ocean characteristics and the presence and size of fish aggregations associated with DFADs, enabling the study of pelagic communities. This unprecedented amount of data is characterized by ocean-scale coverage with high spatial and temporal resolutions, but also by different data formats and specifications depending on buoy model and brand, as well as on the type of data exchange agreements into play. Their use for scientific and management purposes is therefore critically dependent on the abilities of algorithms to process heterogeneous data formats and resolutions. This paper proposes a unified set of algorithms for processing the buoys location data used by the two major purse seine fleets operating in the Atlantic and Indian oceans. Three main issues that need to be addressed prior to the exploitation of the data are identified (structural errors, data records on land and on-board vessels) and five specific filtering criteria are proposed to improve the data cleaning process and, hence, quality. Different filtering procedures are also compared, and their advantages and limitations are discussed.