Seasonality of North Atlantic phytoplankton from space: impact of environmental forcing on a changing phenology (1998–2012)

Abstract

Seasonal pulses of phytoplankton drive seasonal cycles of carbon fixation and particle sedimentation, and might condition recruitment success in many exploited species. Taking advantage of long-term series of remotely sensed chlorophyll a (1998-2012), we analysed changes in phytoplankton seasonality in the North Atlantic Ocean. Phytoplankton phenology was analysed based on a probabilistic characterization of bloom incidence. This approach allowed us to detect changes in the prevalence of different seasonal cycles and, at the same time, to estimate bloom timing and magnitude taking into account uncertainty in bloom detection. Deviations between different sensors stressed the importance of a prolonged overlap between successive missions to ensure a correct assessment of phenological changes, as well as the advantage of semi-analytical chlorophyll algorithms over empirical ones to reduce biases. Earlier and more intense blooms were detected in the subpolar Atlantic, while advanced blooms of less magnitude were common in the Subtropical gyre. In the temperate North Atlantic, spring blooms advanced their timing and decreased in magnitude, whereas fall blooms delayed and increased their intensity. At the same time, the prevalence of locations with a single autumn/winter bloom or with a bimodal seasonal cycle increased, in consonance with a poleward expansion of subtropical conditions. Changes in bloom timing and magnitude presented a clear signature of environmental factors, especially wind forcing, although changes on incident photosynthetically active radiation and sea surface temperature were also important depending on latitude. Trends in bloom magnitude matched changes in mean chlorophyll a during the study period, suggesting that seasonal peaks drive long-term trends in chlorophyll a concentration. Our results link changes in North Atlantic climate with recent trends in the phenology of phytoplankton, suggesting an intensification of these impacts in the near future.

The predictive skill of species distribution models for plankton in a changing climate

“…As PPV has rarely been estimated for SDM predictions (Liu et al, 2009), we unfortunately cannot directly compare these numbers with those of other taxa. Nevertheless, PPV for plankton may be particularly limited: plankton, and in particular phytoplankton, are short-lived organisms which undergo distinct phases of boom and bust (Mackas et al, 2012;Gonz alez Taboada & Anad on, 2014), creating local patchiness. Ocean currents and eddies lead to further local structuring of the pelagic environment and disperse plankton patches laterally beyond suitable areas (Barton et al, 2010).…”

Impact of missing data on the estimation of ecological indicators from satellite ocean-colour time-series

“…The selected indicators are key to characterise and monitor the composition, structure and functioning of the marine ecosystem on seasonal, interannual, decadal and longer time-scales. Thus, an evaluation of the confidence range in the estimates is essential, especially for detection of trends influenced by large-scale environmental and climate drivers (González Taboada and Anadón, 2014;Martinez, Antoine, D'Ortenzio, & De Boyer Montégut, 2011;Racault et al, 2012;Thomalla et al, 2011;Vantrepotte & Mélin, 2009;Zhai et al, 2010;Zhai, Platt, Tang, Sathyendranath, & Walne, 2013).…”

“…The distribution of missing data in the CZCS time-series has been evaluated at monthly resolution (Antoine et al, 2005). However, monthly resolution is not sufficient to assess inter-annual variability and trends in phytoplankton phenology, which are driven by natural or anthropogenic forcing (Chiba et al, 2008;González Taboada & Anadón, 2014;Racault, Le Quéré, Buitenhuis, Sathyendranath, & Platt, 2012;Thomalla, Fauchereau, Swart, & Monteiro, 2011).…”

Winter‐mixing preconditioning of the spring phytoplankton bloom in the Bay of Biscay

“…If no autumn bloom was observed (i.e., only one [Chl a ] SAT maximum occurred), then spring bloom rise was defined as the date when the [Chl a ] SAT seasonal curve increased at the fastest rate. This criterion to identify the timing of the spring bloom avoids the use of an a priori chlorophyll threshold (for references discussing different criteria, see Brody et al ; Blondeau‐Patissier et al ; González Taboada and Anadón ). We defined bloom decay as the day when the [Chl a ] SAT seasonal curve decreased at the fastest rate after both the spring bloom maximum (i.e., the magnitude of the spring bloom peak, max Chl a ) and minimum SST.…”

“…The North Atlantic spring phytoplankton bloom is the most pronounced bloom in open ocean waters (Yoder et al ), although its characteristics vary substantially in space and time (Ueyama and Monger ; Racault et al ; González Taboada and Anadón ). Interannual changes in the timing and magnitude of phytoplankton blooms can lead to a trophic match‐mismatch that modulates the survival of upper trophic levels, including commercially fished stocks (Cushing ; Platt et al ; Durant et al ; Koeller et al ; Kristiansen et al ).…”