Influence of bloom dynamics on Particle Export Efficiency in the North Atlantic: a comparative study of radioanalytical techniques and sediment traps

Highlights: • Performances of 3 methods to obtain POC flux and export efficiency are examined. • Bloom dynamics and sampling date must be associated to the method time scale. • Discrepancies between PEALAGRA, 234Th and 210Po methods are satisfactory interpreted. • Seasonal variability of the export efficiency in the North Atlantic is detected. Abstract: The Biological Carbon Pump is an important component of the global carbon cycle is (BCP). Particle Export Efficiency (PEeff), defined as the proportion of primary production (PP) exported as Particulate Organic Carbon (POC) from the surface ocean, is increasingly used as a metric of the strength of the BCP. However our knowledge of which factors drive variability of PEeff remains poor. This is partially because comparisons of PEeff in different regions often overlook the timescale over which the method used operates in relation to the phase of the plankton bloom. Here we use three techniques to estimate PEeff in situ in the North Atlantic: the radioactive pairs 238U-234Th and 210Pb-210Po, and neutrally buoyant sediment traps (PELAGRA). Order of magnitude discrepancies between values of PEeff obtained from PELAGRA relative to those obtained when applying both radionuclide techniques. POC export fluxes and satellite-derived PP suggest that this results from the differing time scales covered by the three methods and the timing of observations relative to the bloom peak. None of the three techniques are considered inappropriate to estimate PEeff in situ, but bloom dynamics must be considered in relation to the duration over which a particular sampling method operates. Our results suggest a strong seasonal variability in PEeff, most likely controlled by the community structure and hydrographic conditions. This implies that the methods used (specifically their inherent timescales) and the phase of the bloom at the time of sampling must be carefully taken into account to ensure that individual PEeff estimates compiled from different sources to construct global export algorithms are comparable.