Impact of particle aggregation on vertical fluxes of organic matter

https://doi.org/10.1016/j.pocean.2009.07.047Get rights and content

Abstract

Particles sinking out of the euphotic zone are important vehicles of carbon export from the surface ocean. Most of the particles produce heavier aggregates by coagulating with each other before they sink. We implemented an aggregation model into the biogeochemical model of Regional Oceanic Modelling System (ROMS) to simulate the distribution of particles in the water column and their downward transport in the Northwest African upwelling region. Accompanying settling chamber, sediment trap and particle camera measurements provide data for model validation. In situ aggregate settling velocities measured by the settling chamber were around 55 m d−1. Aggregate sizes recorded by the particle camera hardly exceeded 1 mm. The model is based on a continuous size spectrum of aggregates, characterised by the prognostic aggregate mass and aggregate number concentration. Phytoplankton and detritus make up the aggregation pool, which has an averaged, prognostic and size dependent sinking. Model experiments were performed with dense and porous approximations of aggregates with varying maximum aggregate size and stickiness as well as with the inclusion of a disaggregation term. Similar surface productivity in all experiments has been generated in order to find the best combination of parameters that produce measured deep water fluxes. Although the experiments failed to represent surface particle number spectra, in the deep water some of them gave very similar slope and spectrum range as the particle camera observations. Particle fluxes at the mesotrophic sediment trap site off Cape Blanc (CB) have been successfully reproduced by the porous experiment with disaggregation term when particle remineralisation rate was 0.2 d−1. The aggregation–disaggregation model improves the prediction capability of the original biogeochemical model significantly by giving much better estimates of fluxes for both upper and lower trap. The results also point to the need for more studies to enhance our knowledge on particle decay and its variation and to the role that stickiness play in the distribution of vertical fluxes.

Introduction

Particle fluxes from the ocean’s surface layers to its bottom are important means of carbon export and therefore have a crucial role in the global regulation of atmospheric CO2. Many particles sinking into deep layers do so in the form of aggregates, which are produced by the coagulation of smaller particles. Coagulation does not only enhance the removal of material from euphotic zone, but it can also control the maximum phytoplankton concentration in the ocean (Jackson and Kiørboe, 2008). Among eastern boundary upwelling systems, the NW African Upwelling retains elevated concentrations of aggregates due to high biological productivity, is distinguished by their enhanced export and belongs to the most productive ecosystems of the world’s ocean.

The efficiency of aggregate formation is regulated by factors like particle concentration, stickiness and collision rate. Although settling rates generally increase with aggregate size (Alldredge and Gotschalk, 1988, Alldredge and Gotschalk, 1989), large size does not always guarantee fast sinking (Kiørboe et al., 1998) and sinking characteristics of aggregates depend on their composition, shape and porosity. The processes that can reduce the aggregate size are remineralisation, zooplankton feeding (Stemmann et al., 2004), fragmentation by zooplankton (Dilling and Alldredge, 2000) and disaggregation due to physical shear, which is caused by turbulence. The processes that aggregates go through during their sedimentation determine their vertical flux and distribution of elements they carry in the water column.

A number of models have been developed for a numerical definition of aggregation processes in the ocean with different levels of complexity. According to how collision between particles is described, Jackson (2005) classifies these models into two groups as those having rectilinear and curvilinear coagulation kernels. Rectilinear approximation belongs to the early formulations of coagulation, which assume that particles do not influence water motion and that chemical attraction and repulsion between particles can be ignored. Curvilinear kernels, on the contrary, take into account the effect of the particles on the surrounding fluid and on each other, i.e. van der Waals forces, and predict significantly less collision between particles of dissimilar sizes (Jackson and Burd, 1998). Rectilinear kernels appear, therefore, to generate a much stronger coagulation effect and vertical flux (Jackson, 2001).

Despite being the dominant process in controlling vertical carbon flux from biological production (Jackson et al., 2005) there has been almost no implementation of aggregation models on 3-D oceanic-biogeochemical models. Gehlen et al. (2006) presented a first attempt by employing Kriest’s (2002) model to extend the simple parameterisation of aggregation used in the global biogeochemical model of Aumont and Bopp (2006) into its detritus pool. Dadou et al. (2001) developed a 1-D biogeochemical model to reproduce organic matter fluxes in the NW African Upwelling zone and parameterised aggregation as a second-order process. Herein we use a biogeochemical model in a regional setup of the NW African Upwelling System and consider both phytoplankton and detritus as the constituents of aggregates. The model response to changes in porosity, maximum aggregate size and stickiness is evaluated. The purpose is not to detect which aggregate composition (dense or porous) represents the particle behaviour in the ocean, but rather to find out which set of parameters like aggregate density, i.e. settling velocity, and remineralisation rate give the best estimate of deep fluxes with the same or similar surface productivity pattern. We then compare the simulated particle number concentrations and the particle size spectra with those as measured by the camera profiles. The predicted organic carbon fluxes are also compared to the recordings of sediment traps located off Cape Blanc as well as to the fluxes produced by the original biogeochemical model. By bringing the findings from the field measurements like particle camera and settling chamber together with modelling work this paper tries to shed a little light upon the complexities surrounding the particle transformations in the water column.

Section snippets

Sediment trap, particle camera and settling chamber measurements

We refer to Fischer and Karakaş (2009), for the sediment trap characteristics and analysis of the collection. The particle camera (ParCa) system we used was a modified version of the system described in Ratmeyer and Wefer (1996). It consisted of a NIKON Coolpix digital camera with a 3.34 megapixel resolution. A strobe mounted perpendicular to the camera illuminated the water. Each image recorded the particles within a volume of 12.4 l. A SeaBird 36 telemetry unit and a SeaBird 19 CTD (equipped

Comparison of surface chlorophyll fields to the satellite data

The validation of the generated physical fields by the model has been discussed in Karakaş et al., 2006, Marchesiello and Estrade, 2007. In Fig. 2, we illustrate mean surface chlorophyll distribution from SeaWiFS and model experiments for the year 2002. The general pattern of chlorophyll distribution is similar in all simulations to the satellite data and there is very little discrepancy. This is not surprising because by setting the remineralisation rate proportional to the sinking velocity

Summary and conclusions

Various model experiments were carried out by applying an aggregation model into the phytoplankton and detritus compartments of a biogeochemical model. The model was based on a continuous size spectrum of aggregates. The simulations covered maximum aggregate sizes ranging between 0.4 and 5 mm. Both porous and dense aggregates were considered with reduced stickiness in order to yield limited collision between particles. One experiment also involved a disaggregation term to reduce the size of

Acknowledgements

We thank anonymous referees for their constructive comments that helped to improve the manuscript significantly. The numerical experiments were partly carried out on IBM pSeries 690 Supercomputer of Norddeutscher Verbund für Hoch- und Höchstleistungsrechnen (HLRN). This research was funded by the German Research Foundation (DFG) – Centre for Marine Environmental Sciences (MARUM).

References (44)

  • I. Kriest

    Different parameterizations of marine snow in a 1D-model and their influence on representation of marine snow, nitrogen budget and sedimentation

    Deep-Sea Research Part I

    (2002)
  • I. Kriest et al.

    Representing phytoplankton aggregates in biogeochemical models

    Deep-Sea Research Part I

    (1999)
  • J.H. Martin et al.

    VERTEX: carbon cycling in the northeast Pacific

    Deep-Sea Research

    (1987)
  • I.N. McCave

    Size spectra and aggregation of suspended particles in the deep ocean

    Deep-Sea Research

    (1984)
  • U. Passow

    Transparent exopolymer particles (TEP) in aquatic environments

    Progress in Oceanography

    (2002)
  • P. Penven et al.

    Software tools for pre- and post-processing of oceanic regional simulations

    Environmental Modelling and Software

    (2008)
  • A.F. Shchepetkin et al.

    The regional oceanic modeling system (ROMS): A split-explicit, free-surface, topography-following-coordinate oceanic model

    Ocean Modelling

    (2005)
  • L. Stemmann et al.

    A vertical model of particle size distributions and fluxes in the midwater column that includes biological and physical processes – part I: model formulation

    Deep-Sea Research Part I

    (2004)
  • A.L. Alldredge et al.

    In situ settling behaviour of marine snow

    Limnology and Oceanography

    (1988)
  • J.I. Antonov et al.

    World Ocean Atlas 2005

  • O. Aumont et al.

    Globalizing results from ocean in situ iron fertilization studies

    Global Biogeochemical Cycles

    (2006)
  • W.M. Berelson

    Particle settling rates increase with depth in the ocean

    Deep-Sea Research Part II

    (2002)
  • Cited by (31)

    • Transport of organic-walled dinoflagellate cysts in nepheloid layers off Cape Blanc (N-W Africa)

      2018, Deep-Sea Research Part I: Oceanographic Research Papers
      Citation Excerpt :

      Especially a INL at ca. 250–600 m water depth and a BNL could be detected during several years of observation (Fischer et al., 2009; Basse et al., 2014; Ohde et al., 2015 and references therein). INLs in the Cape Blanc region are considered to be particle plumes coming from pulse like resuspension of particles from the sediments at shelf breaks of which some of them are considered to be initiated by bottom trawling activities (e.g. Karakas et al., 2009, Fischer et al., 2009; Ohde et al., 2015). BNL’s are often related with the presence of density gradients and of long internal solitary waves (Bourgault et al., 2014).

    • The influence of extracellular polysaccharides, growth rate, and free coccoliths on the coagulation efficiency of Emiliania huxleyi

      2015, Marine Chemistry
      Citation Excerpt :

      The coagulation efficiency (α) is defined as the fraction of total particle collisions that result in particles attaching to one another and has also been referred to as stickiness (Kiørboe et al., 1990; Engel, 2000), stickiness efficiency (Kahl et al., 2008), sticking coefficient (Logan et al., 1995), coalescence efficiency (McCave, 1984), and collision efficiency factor (Gibbs, 1982). This value has been used in aggregation models to interpret relative changes in carbon export over the duration of a bloom (Jackson and Lochmann, 1993; Kahl et al., 2008; Karakas et al., 2009). In a study of diatoms, Kahl et al. (2008) found that the coagulation efficiency was dependent on the physiological state of the cells, and that this could have a significant effect on the critical concentration of algal cells, and in turn the subsequent export flux.

    View all citing articles on Scopus
    View full text