RESEARCH ARTICLES Eddy-driven stratification initiates North Atlantic spring phytoplankton blooms

Woods Hole Oceanographic Institution, Woods Hole, MA 02543, USA.
Science (Impact Factor: 33.61). 07/2012; 337(6090):54-8. DOI: 10.1126/science.1218740
Source: PubMed


Springtime phytoplankton blooms photosynthetically fix carbon and export it from the surface ocean at globally important rates.
These blooms are triggered by increased light exposure of the phytoplankton due to both seasonal light increase and the development
of a near-surface vertical density gradient (stratification) that inhibits vertical mixing of the phytoplankton. Classically
and in current climate models, that stratification is ascribed to a springtime warming of the sea surface. Here, using observations
from the subpolar North Atlantic and a three-dimensional biophysical model, we show that the initial stratification and resulting
bloom are instead caused by eddy-driven slumping of the basin-scale north-south density gradient, resulting in a patchy bloom
beginning 20 to 30 days earlier than would occur by warming.

72 Reads
  • Source
    • "In the ocean, resolution finer than 10 km is required to simulate the narrow continental shelves where many productive fisheries are located and where the impacts of acidification may be most intense[Gruber et al., 2012], as well as to capture some of the impacts of mesoscale oceanic eddies. Even finer ocean resolution is required to capture small-scale dynamical structures that have been shown to have an important role in modulating the seasonal cycle[Mahadevan et al., 2012]. However, ''comprehensive'' ocean biogeochemical modules such as those used by the major modeling centers increase the amount of resources required by the ocean component by an amount that can make high-resolution simulations unaffordable. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Earth System Models increasingly include ocean biogeochemistry models in order to predict changes in ocean carbon storage, hypoxia, and biological productivity under climate change. However, state-of-the-art ocean biogeochemical models include many advected tracers, that significantly increase the computational resources required, forcing a trade-off with spatial resolution. Here, we compare a state-of-the art model with 30 prognostic tracers (TOPAZ) with two reduced-tracer models, one with 6 tracers (BLING), and the other with 3 tracers (miniBLING). The reduced-tracer models employ parameterized, implicit biological functions, which nonetheless capture many of the most important processes resolved by TOPAZ. All three are embedded in the same coupled climate model. Despite the large difference in tracer number, the absence of tracers for living organic matter is shown to have a minimal impact on the transport of nutrient elements, and the three models produce similar mean annual preindustrial distributions of macronutrients, oxygen, and carbon. Significant differences do exist among the models, in particular the seasonal cycle of biomass and export production, but it does not appear that these are necessary consequences of the reduced tracer number. With increasing CO2, changes in dissolved oxygen and anthropogenic carbon uptake are very similar across the different models. Thus, while the reduced-tracer models do not explicitly resolve the diversity and internal dynamics of marine ecosystems, we demonstrate that such models are applicable to a broad suite of major biogeochemical concerns, including anthropogenic change. These results are very promising for the further development and application of reduced-tracer biogeochemical models that incorporate "sub-ecosystem-scale" parameterizations.
    Full-text · Article · Nov 2015 · Journal of Advances in Modeling Earth Systems
    • "Taylor andconjecture that the increase of primary production induced by such an earlier start of spring bloom at high latitude fronts likely increases the ocean uptake of carbon dioxide and plays an important role in the global carbon cycle. In the absence of slumping of isopycnals by MLIs (as would be the case in our experiments at 10 km resolution), the simulated bloom only appears after the restratification is made by surface heat fluxes, inducing a delay up to 30 days in the emergence of the bloom (Mahadevan et al., 2012). This 30 day delay is also seen in our simulations between the 2 km and 10 km experiments considering the mixed layer restratification, so we would expect the same 30 day delay in the emergence of a bloom if we had a biogeochemical model plugged into our model. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Recent realistic high resolution modeling studies show a net increase of submesoscale activity in fall and winter when the mixed layer depth is at its maximum. This submesoscale activity increase is associated with a reduced deepening of the mixed layer. Both phenomena can be related to the development of mixed layer instabilities, which convert available potential energy into submesoscale eddy kinetic energy and contribute to a fast restratification by slumping the horizontal density gradient in the mixed layer. In the present work, the mixed layer formation and restratification were studied by uniformly cooling a fully turbulent zonal jet in a periodic channel at different resolutions, from eddy resolving (10 km) to submesoscale permitting (2 km). The effect of the submesoscale activity, highlighted by these different horizontal resolutions, was quantified in terms of mixed layer depth, restratification rate and buoyancy fluxes. Contrary to many idealized studies focusing on the restratification phase only, this study addresses a continuous event of mixed layer formation followed by its complete restratification. The robustness of the present results was established by ensemble simulations. The results show that, at higher resolution, when submesoscale starts to be resolved, the mixed layer formed during the surface cooling is significantly shallower and the total restratification is almost three times faster. Such differences between coarse and fine resolution models are consistent with the subme- soscale upward buoyancy flux, which balances the convection during the formation phase and accelerates the restratification once the surface cooling is stopped. This submesoscale buoyancy flux is active even be- low the mixed layer. Our simulations show that mesoscale dynamics also cause restratification, but on longer time scales. Finally, the spatial distribution of the mixed layer depth is highly heterogeneous in the presence of submesoscale activity, prompting the question of whether it is possible to parameterize submesoscale effects and their effects on the marine biology as a function of a spatially-averaged mixed layer depth.
    No preview · Article · Oct 2015 · Ocean Modelling
    • "Phytoplankton concentration, a major component of marine biogeochemical cycles (Broecker et al., 1982; Sarmiento and Gruber, 2006), is driven by the availability of nutrients in the euphotic zone (Valiela, 1995). Known to be influenced by climatological processes such as wind speed, eddies and fronts, and pathways of currents (Taylor and Ferrari, 2011; Mahadevan et al., 2012), phytoplankton concentrations play an important role in the variability of oceanic carbon cycles (Thomalla et al., 2011). The principal photosynthetic pigment in phytoplankton, chlorophyll-a (hereafter, Chl-a), is the most widely used indicator for phytoplankton abundance. "
    [Show abstract] [Hide abstract]
    ABSTRACT: The biological variability of the upwelling region of the Seychelles–Chagos Thermocline Ridge (SCTR), both at surface and subsurface levels, is investigated using monthly outputs of a coupled biophysical model from 1958 to 2011. Owing to its large spatial distribution and sensitivity to climate variability, the SCTR is studied as three distinct regions; namely, sub-regions 1 (western; 5°S–12°S, 55°E–65°E), 2 (central; 5°S–12°S, 65°E–75°E) and 3 (eastern; 5°S–12°S, 75°E–90°E). Surface and subsurface chlorophyll-a (Chl-a) exhibit completely different response mechanisms in sub-region 3 compared to sub-regions 1 and 2 during El Niño Southern Oscillation (ENSO) and Indian Ocean Dipole (IOD) events. During the intense 1997/1998 ENSO–IOD event, the high Chl-a tongue observed in the eastern Indian Ocean induces an increase in surface concentration in sub-region 3, whose subsurface variability is also substantially less (more) impacted by downwelling (upwelling) Rossby waves generated by El Niño (La Niña) forcing. After filtering out the annual signal, wavelet analysis of surface Chl-a revealed a significant 6 month periodicity in sub-regions 1 and 2 whereas a 5-year signal dominated in sub-region 3. The latter suggests that sub-region 3 is more prone to different ENSO/IOD influences, due to its proximity to the eastern Indian Ocean. In the unfiltered data, the subsurface Chl-a in sub-region 3 exhibits a strong signal near 1 year, with sub-regions 1 and 2 having a pronounced 6-year and 5-year signals respectively. These analyses show that the SCTR cannot be investigated as a single homogeneous region due to its large spatial distribution and different response mechanisms to climate events. Furthermore, changes in SST, thermocline depth, winds and Chl-a before and after the 1976–1977 climate shift differed across the SCTR, further highlighting the heterogeneity of this sensitive region in the Indian Ocean.
    No preview · Article · Oct 2015 · Journal of Marine Systems
Show more