Abstract
Marine aggregates formed through particle coagulation, large ones (>0.05 cm) also called marine snow, make a significant contribution to the global carbon flux by sinking from the euphotic zone, impacting the Earth’s climate. Since aggregate sinking velocity and carbon content are size-dependent, understanding the physical mechanisms controlling aggregate size distribution is fundamental to determining the biological carbon pump efficiency. Theoretical, laboratory and in-situ studies of flocculation have suggested that turbulence in the benthic boundary layer is important for aggregate formation and destruction, but the small number of field observations has limited our understanding of the role of turbulence on aggregation processes in the ocean surface layer away from energetic boundaries. Using simultaneous field observations of turbulence and aggregates, we show how aggregate formation, destruction, morphology and size distribution in the ocean surface layer (10–100 m) are mediated by interactions between turbulence and aggregate concentration. Our findings suggest that turbulence enhances aggregate formation up to a critical turbulent kinetic energy dissipation rate of 10−6 (W kg−1), above which the smallest turbulent eddies limit aggregate size.
Similar content being viewed by others
Introduction
Marine aggregates range in size from approximately 1 µm to several centimetres1. Formed in the sunlit upper layers of the ocean and composed predominantly of organic material1,2,3, aggregates sink as a constant drizzle to the deep ocean3, exporting energy and acting as hotspots of microbial activity and biogeochemical transformations along the way4. Aggregate formation occurs through the collision and adhesion of smaller particles into larger particles and is driven by three main physical processes5,6. Brownian motion controls the collision of small particles (<1 µm). Differential sinking involves faster settling particles overtaking and colliding with slower settling particles and dominates for particle sizes between about 1 and 100 µm. Turbulent shear dominates the interactions between larger particles5 (>100 µm). Since large aggregates have increased sinking velocity and carbon content relative to small aggregates7, the mechanisms controlling aggregate size distribution in the upper ocean have important consequences for determining the transport of carbon to the deep ocean8,9.
The role of turbulence on aggregates has been investigated theoretically and experimentally10 over a range of flow conditions and materials, usually using idealized shear models and uniform spheres as source particles11,12. However, our understanding of the influence of turbulence on aggregates in the upper ocean interior has remained constrained due to a lack of direct observations. Early laboratory13 and modelling14 studies indicated that turbulence was important for both aggregation and disaggregation, consistent with the conceptual view of sedimentation processes in estuarine systems15. Subsequent observations made in the sediment-rich benthic boundary layer revealed that our understanding of particle disaggregation remains uncertain16 with minimal influence of turbulence detected on aggregate size, even though modelling of disaggregation processes in the bottom boundary layer predicted aggregate breakup under strong turbulence14.
In comparison to early experimental and modelling studies10,11,12,13,14,15,16,17, a laboratory study18 using natural aggregates collected from the ocean’s surface layer (10–15 m depth) found that turbulent kinetic energy dissipation rates as strong as 10−4 W kg−1 did not cause aggregate breakup. Estimates of drag forces on falling aggregates19 have further suggested that sinking-induced stresses may be more effective in causing aggregate breakup than turbulence. The conclusion from these studies18,19 was that disaggregation by turbulence was relatively unimportant in the upper ocean. Experimental11 and numerical studies12 investigating the collision of small particles (<~100 µm) have since shown that turbulence initially enhances aggregation, but disaggregation becomes increasingly important in controlling aggregate size distribution as the system ages and aggregates grow. More recent observations of aggregates made in the coastal benthic boundary layer20,21 and energetic tidal channels22,23 have shown that turbulence may indeed cause aggregate breakup, thereby limiting the size distribution of aggregates formed, reinstating the likely importance of disaggregation by turbulence.
The general applicability of these studies to understanding aggregation processes in the upper layers of the ocean remains uncertain for several reasons. First, the biological composition of marine aggregates that affects aggregation and disaggregation processes is known to be highly sensitive to changes in the ambient environmental conditions and the methods used for collection24. For example, the stickiness of diatoms following collection from the field has been shown to increase due to nutrient and light limitation25,26,27. Aggregates become compacted under even most gentle collection methods, potentially leading to stronger bonds24, which may be the reason aggregates survived the strong turbulence generated in the laboratory experiments18. Aggregates in the benthic boundary layer tend to be much smaller than the marine aggregates observed in the upper ocean water column, in part because their composition contains a higher density of the minerals and sediments28. In comparison, the composition of aggregates formed in the ocean surface layer contains an increased fraction of organic material, including living and dead phytoplankton1,2, fecal pellets3 and extracellular polymeric substances (EPS)29.
Another important difference between laboratory experiments, energetic coastal environments (e.g., bottom boundary layer, tidal channels) and the upper water column of the open ocean is the intensity of turbulence. Turbulent kinetic energy dissipation rates in the upper water column rarely exceed 10−6 W kg−1, except in the top few meters of surface layer when breaking surface waves30,31 generate strong turbulence. Similarly, kinetic energy dissipation rates can far exceed 10−5 W kg−1 in laboratory experiments and in the bottom boundary layer of coastal environments, where waves and currents can generate intense near-bed turbulence30. As a result, the relationship observed between turbulence and aggregates in highly localized bottom boundary layers20,21 and energetic coastal waters22,23 are not likely to be representative of processes occurring in the water column interior of the upper ocean that occupies most of the world ocean.
These uncertainties support the necessity of field measurements in the upper ocean to develop our understanding of the relationship between turbulence and aggregates and its implications for the biological pump under climate change32,33. In the present study, we collected simultaneous measurements of turbulence and aggregates in the upper ocean (~10–100 m) away from energetic coastal environments. We explore how aggregate size and other related properties, such as morphology and volume concentration, are affected by turbulence in the sunlit upper layer of the world ocean where particles are formed by primary production.
Methods
We made non-disruptive measurements of turbulence and aggregates in the upper ocean water column, between the surface and 100 m. Measurements were made during 10 campaigns and multiple seasons in coastal and offshore waters of Japan.
Microscale variations in temperature and turbulent velocity were measured with a free-fall microstructure profiler (TurboMAP-L, JFE Advantech Co., Ltd.)34 at a sample rate of 512 Hz and fall-speed of ~0.5 m s−1. We estimated the turbulent kinetic energy dissipation rate (ε, W kg−1) by integrating the turbulent velocity shear spectrum obtained from the shear probe over 2 second segments (~ 1 m) from approximately 1 cycle per meter to half the Kolmogorov wavenumber ((ν3/ε)−1/4)35,36, where ν is the kinematic viscosity of seawater. A correction was made to recover the unresolved variance37 using the Nasmyth empirical spectrum38. The size of the shear probe that measures turbulent velocity was designed to resolve an expected minimum level of ε ~ 10−10 W Kg−139 under the assumption of isotropic turbulence40. Although turbulence may not be isotropic when ε is low, axisymmetric turbulence theory that accounts for stratification effects on turbulence indicates the error associated with use of the isotropic turbulence theory is less than 35%41. Therefore, ε estimates based on isotropic turbulence theory are a reasonable approximation to its true value. To avoid contamination by vessel-generated turbulence, we discarded ε observations obtained within 10 m of the surface. Increases in the 1 m scale turbidity or ε were used to detect the presence of bottom boundary layer in waters less than 100 m deep. Since these signals were typically detected much closer than 10 m from the bottom, we discarded observations made within the bottom 10 m of all profiles to avoid contaminating water column observations with those from benthic boundary layers.
A mini CMOS camera (DSL II 190, Little Leonard Inc.)34 mounted on TurboMAP-L collected images of aggregates at a sampling rate of 5 Hz simultaneously with shear observations. Processed images had a field of view of 2 cm × 2 cm and a pixel resolution of 59 µm34. Streaked images were identified by assessing the 2D image spectrum using a 2D Fast Fourier Transform (2D FFT)34,42. The 2D spectrum is a symmetric circle when images are not smeared, whereas asymmetry is seen in the 2D spectrum of streaked images. To test for asymmetry, two perpendicular sets of 1D spectra were chosen and the ratio of variance for each wavenumber was calculated for each perpendicular pair. The variance ratio is approximately 1 in unstreaked images, with images rejected from further analysis if the average variance ratio for one perpendicular pair exceeded 1.5 or 1/1.5. This criterion assesses smearing across all aggregates imaged in the field of view and minimizes the rejection of images which contain rare individual long and thin aggregates.
Individual aggregates were then approximated as ellipses using the regionprops function in MATLAB (Mathworks Inc.) to determine major (MajAL) and minor (MinAL) axis lengths and equivalent spherical diameters (ESD). To focus on the larger size fraction of aggregates expected to be influenced by turbulence5, only objects with MajAL > 0.03 cm were considered to be aggregates. Coincident high-resolution fluorescence microstructure profiling that resolved millimetre scale changes in chlorophyll-a fluorescence showed extremely strong signals where aggregates were seen34, implying that aggregates captured by the DSL camera contained live phytoplankton. Additional laboratory tank experiments using particles of known size were also conducted to confirm that unfocused particles and streaked images were removed by the size threshold and 2D spectrum criteria. In total, 57,669 images collected over 148 profiles were retained. A total of 1,269,978 aggregates were identified; among them 1,103,412 aggregates were observed for ε < 10−6 W kg−1 and 166,566 aggregates for ε > 10−6 W kg−1.
Relationships between turbulence and aggregates were then examined using 10 m scale average properties that included, the average turbulent kinetic energy dissipation rate (\(\bar{\varepsilon }\), W kg−1), total aggregate volume concentration (Vagg, ppm), aggregate minor axis length (\(\bar{{M}{i}{n}{A}{L}}\), cm), major axis length (\(\bar{{M}{a}{j}{A}{L}}\), cm), equivalent spherical diameter (\(\bar{{E}{S}{D}}\), cm) and aspect ratio (\(\bar{{A}{R}}\,=\,\frac{\bar{{M}{a}{j}{A}{L}}}{\bar{{M}{i}{n}{A}L}}\)), where the over bar represents the 10 m scale mean value. Since the imaging system provides a 2D image of a 3D object, differences in aggregate size due to orientation are expected to be reduced by the use of 10 m scale average metrics. The volume of an individual aggregate is calculated as \(\frac{\pi }{{6}}{{E}{S}{D}}^{3}\), where Vagg is the fraction of volume occupied by aggregates and is expressed in cm3 m−3, equivalent to parts per million (ppm).
Aggregate number spectra43 (n) were used to describe the size distribution of aggregates. The number of aggregates (ΔN) in logarithmically increasing MajAL bins of average size d was divided by the bin width (∆d) and the sample volume to construct a number spectrum. Any ΔN < 10 was discarded before computing n. For each spectrum, a bilinear relationship was fit to log(n) as a function of log(d) to obtain values for the inflection point (Lint) and the slope below (slope 1, small aggregates) and above (slope 2, large aggregates) the Lint. The mean sum of squared error of each fit was then calculated. Different Lint’s were then selected at intervals of ∆d either side of the first Lint and the slopes determined. The final accepted Lint and slopes were those with the smallest mean sum of squared error.
Finally, the distribution of aggregate volume as a function of the ESD size expressed as the normalised volume distribution nVd43,44 was estimated for each order of magnitude of \(\bar{\varepsilon }\) of between 10−10 < o(\(\bar{\varepsilon }\)) < 10−5 W kg−1. For each o(\(\bar{\varepsilon }\)) interval, the number of aggregates in logarithmically increasing ESD bin sizes (d) was divided by the bin width (∆d) and sample volume to construct an ESD number spectrum. ESD number spectra were multiplied by V = \(\frac{\pi }{{6}}{{E}{S}{D}}^{3}\) and d to obtain nVd, whereby the integral of nVd is equivalent to \({V}_{agg}=\int nV\,{\rm{d}}d\,=\int nVd\,{\rm{d}}(ln\,d)\cdot \,\)
Results and Discussion
Values of \(\bar{\varepsilon }\) ranged from 10−10 to 10−5 W kg−1 (Fig. 1a) and spanned the full range of naturally occurring turbulence intensities found in the upper ocean interior, away from energetic surface and bottom boundary layer regions30,31. Aggregate \(\bar{{M}{a}{j}{A}{L}}\) ranged between 0.031 and 0.133 cm and log10 (\(\bar{{M}{a}{j}{A}{L}}\)) and was positively correlated with log10(\(\bar{\varepsilon }\)) (Fig. 1a, r2 = 0.52, n = 567, p < < 0.001). The majority of (\(\bar{{M}{a}{j}{A}{L}}\)) were smaller than the size of the smallest turbulent eddies, here defined by the Kolmogorov length scale (Lk = (ν3/ε)1/4). Positive correlation between log10(\(\bar{{M}{a}{j}{A}{L}}\)) and log10(Vagg) (Fig. 1b, r2 = 0.74, n = 567, p ≪ 0.001) indicates that Vagg is also a crucial factor determining the size distribution of aggregates, since aggregate total volume is a measure of the number of particles available for coagulation6. Higher particle concentrations should increase coagulation rates, leading to larger particles, while sinking and disaggregation prevent particles from becoming indefinitely large. Multiple linear regression analysis showed that log10(\(\bar{\varepsilon }\)) and log10(Vagg) collectively explained 81% of the variance in log10(\(\bar{{M}{a}{j}{A}{L}}\)) (r2 = 0.81, n = 567, p < < 0.01), with log10(\(\bar{\varepsilon }\)) contributing 32% and log10 (Vagg) 68% to this correlation.
Our findings show 97% of \(\bar{{M}{a}{j}{A}{L}}\) values were below 0.1 cm in size (Fig. 1a), which is the size Lk for \(\bar{\varepsilon }\) = 10−6 W kg−1 and the upper limit of dissipation rates typically observed in the ocean interior30,31. The positive correlation observed at length scales smaller than Lk demonstrates that turbulence enhancement of aggregation occurs at higher rate than disaggregation when shear is laminar22,45 and aggregate sizes are smaller than Lk. This results in the net formation of larger aggregates. When the flow scale is equal to the Kolmogorov scale, the Reynolds number is 1 and the flow is very viscous, hence the resulting flow at this length scale is laminar shear40. Previous bottom boundary observations have shown a decrease in aggregates size when Lk was smaller than 0.1 cm, equivalent ε > 10−6 W kg−122. Therefore, we expect that for \(\bar{\varepsilon }\) > 10−6 W kg−1 and \(\bar{{M}{a}{j}{A}{L}} < {L}_{k}\)the disaggregation rate exceeds the aggregation rate, as shear associated with the smallest turbulent eddies causes breakup and inhibits further size increases.
While values of \(\bar{{M}{a}{j}{A}{L}}\) shown in Fig. 1 were calculated using 10 m averages, the number of individual aggregates with MajAL larger than Lk remained relatively small. Above \(\bar{\varepsilon }\) ≥ 10−6 W kg−1, 63% of individual MajAL sampled (non-averaged samples, n = 166,566) were smaller than Lk. At lower turbulent intensities, the proportion of individual aggregates smaller than Lk increased from 80% at \(\bar{\varepsilon }\) = o(10−7 W kg−1) to 99% of aggregates at \(\bar{\varepsilon }\) = o(10−10 W kg−1). This shift demonstrates that aggregate size distribution is a dynamic property, with the potential for some aggregates to increase in size even under high average turbulent intensities (\(\bar{\varepsilon }\) > 10−6 W kg−1) and for others to undergo disaggregation at lower average turbulent intensities (\(\bar{\varepsilon }\) < 10−6 W kg−1). This interpretation is supported by results shown in Fig. 2, which demonstrates increases in the variability of individual aggregate sizes around the mean, expressed as the coefficient of variation, (CVMajAL = standard deviation/mean), under increasing turbulent intensities.
For a more direct comparison between turbulence and aggregate size we calculated the mean size of individual aggregates for each order of magnitude of \(\overline{\,\varepsilon }\)(\(\bar{{{M}{a}{j}{A}{L}}_{\varepsilon }}\), cm). Increases in \(\bar{{{M}{a}{j}{A}{L}}_{\varepsilon }}\) dropped from ~20% between \(\bar{\varepsilon }\) = o(10−10 W kg−1) and o(10−9 W−1) to ~10% between \(\bar{\varepsilon }\) = o(10−7 W kg−1) and o(10−6 W kg−1) (Fig. 2a) and were associated with a corresponding increase in the coefficient of variation (CVMajAL) from 0.69 to 0.98 (Fig. 2b). The plateau in CVMajAL observed at higher turbulent intensities is consistent with disaggregation rates increasing as both turbulence levels and the average size of aggregates increase (Fig. 1a).
Increases in aggregate size with turbulence were also associated with changes in aggregate morphology (Fig. 3). The increase in (\(\bar{{A}{R}})\) with log10(\(\bar{{M}{a}{j}{A}{L}}\)) (Fig. 3a; r2 = 0.45, n = 567, p < < 0.001) and log10 \(\bar{\varepsilon }\) (Fig. 3b; r2 = 0.40, n = 567, p < < 0.001) shows that aggregates became elongated with increases in size and turbulence intensity. Numerical simulations and laboratory experiments46,47,48 have shown that inertial particles in turbulence cluster in regions of high-strain. Our results suggest that larger aggregates become inertial, possibly being strained by shear due to strong turbulence, resulting in elongation. Whilst increased inertial force on larger aggregates may also enhance breakage under strong turbulence, laboratory experiments49 and numerical simulations50 have demonstrated the settling velocity of elongated phytoplankton increase under elevated turbulence. It is possible that aggregate settling velocity increases due to both morphological changes (Fig. 3b) and size increases (Figs 1 and 2) under increasing turbulence up to a critical turbulent intensity, \(\bar{\varepsilon }\) = 10−6 W kg−1.
Aggregate number spectrum (n, cm−4)43 as a function of log10(MajAL) describes the non-averaged size distribution of individual aggregates throughout the water column. Figure 4 shows a spectrum for individual aggregates sampled between 10 and 100 m depth from the Kuroshio extension (37˚04′05″N, 142˚54′36″E). Here, an average dissipation rate εCA was computed from all individual ε to estimate the corresponding mean Kolmogorov scale (Lk,CA, cm) in these near surface waters (Fig. 4, dashed lines). Two slopes were fitted to the number spectrum (Fig. 4, solid lines), with a gradient of −2.72 for the smaller aggregate size range (slope 1) and −4.53 for the larger aggregate size range (slope 2). There was a decrease in the number of aggregates expected by using the line fit to the smaller aggregates for aggregates larger than the intersection (Lint, cm) of the two lines. Here, Lint is 0.16 cm and the Kolmogorov scale based on εCA is Lk,CA = 0.17 cm. The ratio between Lint and Lk,CAis 0.95 and shows the number of aggregates decreases significantly when aggregate size is larger than Lk,CA. This trend was consistent across all campaigns used in this study. The significant decrease in n as the aggregate size exceeds Lk,CA suggests that the role of particle collision in aggregate formation becomes smaller as disaggregation due to turbulence becomes more prominent.
The normalized volume distribution (nVd, ppm)43,44 as a function of log10(ESD) provides further insight into aggregation and disaggregation processes (Fig. 5). The shape of the nVd distributions is similar to the lognormal distributions described previously43. A simulation51 showed that nVd for aggregates have a lognormal-like distribution when both aggregation and disaggregation are taken into account, suggesting that disaggregation occurs at all level of turbulence (Fig. 5). This is consistent with the original breakage model proposed by Kolmogorov52. Lognormal turbulence theory52,53 shows that a fraction of the water over which \(\bar{\varepsilon }\) is calculated contains localized regions of the turbulent kinetic energy dissipation rate higher than ensemble average54. Hence, for the range of observed \(\bar{\varepsilon }\), parcels of highly localized turbulence may exceed ε = 10−6 W kg−1 and are expected to cause disaggregation even under conditions of low average dissipation rates. The area under the curve of nVd is proportional to Vagg43,44. Increased nVd with \(\bar{\varepsilon }\) is consistent with the aggregation rate increasing under stronger turbulence. The distribution peak shifted to larger ESD with increasing in \(\bar{\varepsilon }\) by ~15–20% when \(\bar{\varepsilon }\) increased one order of magnitude. The increase was limited to ~0.16 cm when \(\bar{\varepsilon }\) = 10−6 – 10−5 W kg−1. For ESD larger than the distribution peak, negative nVd slopes indicate that loss of the large aggregates by disaggregation counters their production by aggregation; steeper slopes at higher \(\bar{\varepsilon }\) show that the loss becomes more rapid as the turbulence intensity increases. This is consistent with turbulence-induced disaggregation rate overtaking the aggregation rate with increased Vagg.
Conclusions
Our observations provide a comprehensive set of simultaneous measurements of aggregate concentrations as a function of size that resolve the full range of turbulent intensities, \(\bar{\varepsilon }\) = 10−10 – 10−6 W kg−1, found within the upper ocean away from energetic near surface and bottom boundary layers (~10–100 m depth). Although turbulent intensities in coastal environments and near boundary layers can far exceed 10−5 W kg−1, our observed values cover the range of intensities found typically over the majority of worlds upper ocean surface layer30,31. Our direct observations show turbulence enhances aggregation up to ~\(\bar{\varepsilon }\) = 10−6 W kg−1 with greater turbulence intensities cause increasing disaggregation, consistent with laboratory13 and theoretical14,17 studies and the early conceptual view of aggregation dynamics studied in coastal environments15. Since most of the ocean upper water column interior contains \(\bar{\varepsilon }\) < 10−6 W kg−130,31, turbulent mediation of aggregate size and morphology is likely to be an important factor influencing a range of biogeochemical processes, including carbon sequestration. This is because aggregate size and morphology are important determinant factors of settling velocities and carbon flux7,55,microbial abundances56,57 and associated biogeochemical activity through bacterial remineralization1,4,56. As climate change is expected to supress turbulence intensity8 and alter phytoplankton communities32,33 in the euphotic zone, the mediation of aggregates by turbulence may have unexpected consequences for global carbon cycle via the biological pump9,58.
Data availability
All data used in this study are available from the first author and corresponding author (M.T. and H.Y.) upon request to jasmine222mari@gmail.com or hide@kaiyodai.ac.jp.
References
Simon, M., Grossart, H. P., Schweitzer, B. & Ploug, H. Microbial ecology of organic aggregates in aquatic ecosystems. Aquatic Microbial Ecology 28, 175–211, https://doi.org/10.3354/ame028175 (2002).
Silver, M. Marine snow: a brief historical sketch. Limnology and Oceanography Bulletin 24, 5–10, https://doi.org/10.1002/lob.10005 (2015).
Turner, J. T. Zooplankton fecal pellets, marine snow, phytodetritus and the ocean’s biological pump. Progress in Oceanography 130, 205–248, https://doi.org/10.1016/j.pocean.2014.08.005 (2015).
Azam, F. & Long, R. A. Sea snow microcosms. Nature 414, 495–498, https://doi.org/10.1038/35107174 (2001).
McCave, I. N. Size spectra and aggregation of suspended particles in the deep ocean. Deep Sea Research 31, 329–352, https://doi.org/10.1016/0198-0149(84)90088-8 (1984).
Jackson, G. A. A model of the formation of marine algal flocs by physical coagulation processes. Deep Sea Research 37, 1197–1211, https://doi.org/10.1016/0198-0149(90)90038-W (1990).
Alldredge, A. The carbon, nitrogen and mass content of marine snow as a function of aggregate size. Deep Sea Research 45, 529–541, https://doi.org/10.1016/S0967-0637(97)00048-4 (1998).
Passow, U. & Carlson, C. A. The biological pump in a high CO2 world. Marine Ecology Progress Series 470, 249–271, https://doi.org/10.3354/meps09985 (2012).
Stemmann, L., Jackson, G. A. & Ianson, D. 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 51, 865–884, https://doi.org/10.1016/j.dsr.2004.03.001 (2004).
Guseva, K. & Feudel, U. Aggregation and fragmentation dynamics in random flows: From tracers to inertial aggregates. Physical Review E 95(6), 062604, https://doi.org/10.1103/PhysRevE.95.062604 (2017).
Soos, M. et al. Effect of shear rate on aggregate size and morphology investigated under turbulent conditions in stirred tank. Journal of Colloid and Interface Science 319, 577–589, https://doi.org/10.1016/j.jcis.2007.12.005 (2008).
Babler, M. U. et al. Numerical simulations of aggregate breakup in bounded and unbounded turbulent flows. Journal of Fluid Mechanics 766, 104–128, https://doi.org/10.1017/jfm.2015.13 (2015).
Dyer, K. R. & Manning, A. J. Observation of the size, settling velocity and effective density of flocs, and their fractal dimensions. Journal of Sea Research 41, 87–95, https://doi.org/10.1016/S1385-1101(98)00036-7 (1999).
Ruiz, J. & Izquierdo, A. A simple model for the break-up of marine aggregates by turbulent shear. Oceanographica Acta 20, 597–605 (1997).
Dyer, K. R. Sediment processes in estuaries: Future research requirements. Journal of Geophysical Research 94, 14327–14339, https://doi.org/10.1029/JC094iC10p14327 (1989).
Hill, P. S., Voulgaris, G. & Trowbridge, J. H. Controls on floc size in a continental shelf bottom boundary layer. Journal of Geophysical Research 106, 9543–9543, https://doi.org/10.1029/2000JC900102 (2001).
Winterwerp, J. C. A simple model for turbulence induced flocculation of cohesive sediment. Journal of Hydraulic Research 36, 309–326, https://doi.org/10.1080/00221689809498621 (1998).
Alldredge, A. L., Granata, T. C., Gotschalk, C. C. & Dickey, T. D. The physical strength of marine snow and its implications for particle disaggregation in the ocean. Limnology and Oceanography 35, 1415–1428, https://doi.org/10.4319/lo.1990.35.7.1415 (1990).
Hill, P. Controls on floc size in the sea. Oceanography 11, 13–18, https://doi.org/10.5670/oceanog.1998.03 (1998).
Mikkelsen, O. A., Hill, P. S. & Milligan, T. G. Single-grain, microfloc and macrofloc volume variations observed with a LISST-100 and a digital floc camera. Journal of Sea Research 55, 87–102, https://doi.org/10.1016/j.seares.2005.09.003 (2006).
Safak, I., Allison, M. A. & Sheremet, A. Floc variability under changing turbulent stresses and sediment availability on a wave energetic muddy shelf. Continental Shelf Research 53, 1–10, https://doi.org/10.1016/j.csr.2012.11.015 (2013).
Braithwaite, K. M., Bowers, D. G., Nimmo Smith, W. A. M. & Graham, G. W. Controls on floc growth in an energetic tidal channel. Journal of Geophysical Research: Oceans 117, 1–12, https://doi.org/10.1029/2011JC007094 (2012).
Bowers, D. G., Binding, C. E. & Ellis, K. M. Satellite remote sensing of the geographical distribution of suspended particle size in an energetic shelf sea. Estuarine, Coastal and Shelf Science 73, 457–466, https://doi.org/10.1016/j.ecss.2007.02.005 (2007).
Alldredge, A. L. & Gotschalk, C. In situ settling behavior of marine snow. Limnology and Oceanography 33, 339–335, https://doi.org/10.4319/lo.1988.33.3.0339 (1988).
Kiørboe, T., Andersen, K. P. & Dam, H. G. Coagulation efficiency and aggregate formation in marine phytoplankton. Marine Biology 107, 235–245, https://doi.org/10.1007/bf01319822 (1990).
Kiørboe, T. Colonization of marine snow aggregates by invertebrate zooplankton: abundance, scaling, and possible role. Limnology and Oceanography 45, 479–484, https://doi.org/10.4319/lo.2000.45.2.0479 (2000).
Engel, A. The role of transparent exopolymer particles (TEP) in the increase in apparent particle stickiness (α) during the decline of a diatom bloom. Journal of Plankton Research 22, 485–497, https://doi.org/10.1093/plankt/22.3.485 (2000).
Larsen, L. G., Harvey, J. W. & Crimaldi, J. P. Morphologic and transport properties of natural organic floc. Water Resources Research 45, https://doi.org/10.1029/2008wr006990 (2009).
Bhaskar, P. V. & Narayan B. Bhosle, N.B. Microbial extracellular polymeric substances in marine biogeochemical processes. Current Science 88(1), 45–53, https://www.jstor.org/stable/24110092 (2005).
Geyer, Wr, Scully, M. E. & Ralston, D. K. Quantifying vertical mixing in estuaries. Environmental Fluid Mechanics 8, 495–509, https://doi.org/10.1007/s10652-008-9107-2 (2008).
Smyth, W. D. & Moum, J. N. 3D turbulence. in Encyclopedia of Ocean Sciences Vol. 3, eds Bokuniewicz, H., Yger, P. & Kirk Cochran, J., 486–496, https://doi.org/10.1016/B978-0-12-409548-9.09728-1 (Academic Press, 2019).
Falkowski, P. G. & Oliver, M. J. Mix and match: how climate selects phytoplankton. Nature reviews microbiology 5(10), 813–819, https://doi.org/10.1038/nrmicro1751 (2007).
Irwin, A. J. et al. Phytoplankton adapt to changing ocean environments. Proceedings of the National Academy of Sciences 112, 5762–5766, https://doi.org/10.1073/pnas.1414752112 (2015).
Doubell, M. J., Yamazaki, H., Li, H. & Kokubu, Y. An advanced laser-based fluorescence microstructure profiler (TurboMAP-L) for measuring bio-physical coupling in aquatic systems. Journal of Plankton Research 31, 1441–1452, https://doi.org/10.1093/plankt/fbp092 (2009).
Doubell, M. J., Prairie, J. C. & Yamazaki, H. Millimeter scale profiles of chlorophyll fluorescence: deciphering the microscale spatial structure of phytoplankton. Deep Sea Research Part II 101, 207–215, https://doi.org/10.1016/j.dsr2.2012.12.009 (2014).
Kokubu, Y., Yamazaki, H., Nagai, T. & Gross, E. S. Mixing observations at a constricted channel of a semi-closed estuary: Tokyo Bay. Continental Shelf Research 69, 1–16, https://doi.org/10.1016/j.csr.2013.09.004 (2013).
Oakey, N. S. & Elliott, J. A. Dissipation Within the Surface Mixed Layer. Journal of Physical Oceanography 12, 171–185, 10.1175/1520-0485(1982) 012<0171:DWTSML>2.0.CO;2 (1982).
Nasmyth, P. Oceanic Turbulence. Ph.D. Thesis, University of British Columbia, 69pp (1970).
Lueck, R. G., Wolk, F. & Yamazaki, H. Oceanic velocity microstructure measurements in the 20th century. Journal of Oceanography 58, 153–174, https://doi.org/10.1023/A:1015837020019 (2002).
Tennekes, H. & Lumley, J. A First Course In Turbulence. (The MIT Press, 1972).
Yamazaki, H. & Osborn, T. Dissipation estimates for stratified turbulence. Journal of Geophysical Research: Oceans 95, 9739–9744, https://doi.org/10.1029/JC095iC06p09739 (1990).
Franks, P. J. S. & Jaffe, J. S. Microscale distributions of phytoplankton: Initial results from a two-dimensional imaging fluorometer, OSST. Marine Ecology Progress Series 220, 59–72, https://doi.org/10.3354/meps220059 (2001).
Petrik, C. M., Jackson, G. A. & Checkley, D. M. Aggregates and their distributions determined from LOPC observations made using an autonomous profiling float. Deep Sea Research 74, 64–81, https://doi.org/10.1016/j.dsr.2012.12.009 (2013).
Jackson, G. A., Checkley, D. M. & Dagg, M. Settling of particles in the upper 100 m of the ocean detected with autonomous profiling floats off California. Deep Sea Research 99, 75–86, https://doi.org/10.1016/j.dsr.2015.02.001 (2015).
Friedlander, S. K. Smoke, dust, and haze: fundamentals of aerosol behavior. (Wiley, New York, 1977).
Eaton, J. K. & Fessler, J. R. Preferential concentration of particles by turbulence. International Journal of Multiphase Flow 20, 169–209, https://doi.org/10.1016/0301-9322(94)90072-8 (1994).
Squires, K. D. & Eaton, J. K. Preferential concentration of particles by turbulence. Physics of Fluids 3, 1169–1178, https://doi.org/10.1063/1.858045 (1991).
de Jong, J. et al. Measurement of inertial particle clustering and relative velocity statistics in isotropic turbulence using holographic imaging. International Journal of Multiphase Flow 36, 324–332, https://doi.org/10.1016/j.ijmultiphaseflow.2009.11.008 (2010).
Ruiz, J., Macias, D. & Peters, F. Turbulence increases the average settling velocity of phytoplankton cells. Proceedings of National Academy of Sciences 101, 17720–17724, https://doi.org/10.1073/pnas.0401539101 (2004).
Clifton, W., Bearon, R. N. & Bees, M. A. Enhanced sedimentation of elongated plankton in simple flows. Journal of Applied Mathematics 83, 743–766, https://doi.org/10.1093/imamat/xxx000 (2018).
Jouandet, M. P. et al. Rapid formation of large aggregates during the spring bloom of Kerguelen Island: observations and model comparisons. Biogeosciences 11, 4393–4406, https://doi.org/10.5194/bg-11-4393-2014 (2014).
Kolmogorov, A. N. On the logarithmical normal particle size distribution caused by particle crushing. Doklady Akademii Nauk SSSR 31, 99–102 (1941).
Gurvich, A. S. & Yaglom, A. M. Breakdown of eddies and probability distributions for small-Scale turbulence. Physics of Fluids 10, 59–65, https://doi.org/10.1063/1.1762505 (1993).
Yamazaki, H. & Squires, K. An application of the lognormal theory to moderate Reynolds number turbulent structures. In Handbook of scaling methods in aquatic ecology measurement, analysis, simulation eds Seuront, L. & Strutton, P. G., Vol. 85 469–478 (CRC Press, 2003).
Laurenceau-Cornec, E. C. et al. The relative importance of phytoplankton aggregates and zooplankton fecal pellets to carbon export: insights from free-drifting sediment trap deployments in naturally iron-fertilised waters near the Kerguelen Plateau. Biogeosciences 12(4), 1007–1027, https://doi.org/10.5194/bg-12-1007-2015 (2015).
Kiørboe, T., Tang, K., Grossart, H. P. & Ploug, H. Dynamics of microbial communities on marine snow aggregates: colonization, growth, detachment, and grazing mortality of attached bacteria. Applied and Environmental Microbiology 69(6), 3036–3047, https://doi.org/10.1128/AEM.69.6.3036-3047.2003 (2003).
Kiørboe, T. Marine snow microbial communities: scaling of abundances with aggregate size. Aquatic Microbial Ecology 33(1), 67–75, https://doi.org/10.3354/ame033067 (2003).
Gehlen, M. et al. Reconciling surface ocean productivity, export fluxes and sediment composition in a global biogeochemical ocean model. Biogeosciences 3, 521–537, https://doi.org/10.5194/bg-3-521-2006 (2006).
Acknowledgements
We thank the captain and the crew of R/T/V Seiyo-maru and all members of Laboratory of Ocean Ecosystems and Dynamics (Tokyo University of Marine Science and Technology) for their support during the field surveys. We thank J. Mitchell and Amatzia Genin for discussions. This study was funded by a Grant-in-Aid for Science Research (B2) 1610005 from Japan Society for the Promotion of Science and Japan Science and Technology Agency CREST Grant Number JPMJCR12A6, Japan.
Author information
Authors and Affiliations
Contributions
H.Y. conceived this study. M.T. and M.D. contributed equally to this work, co-writing the manuscript with input from H.Y. and G.J. Data were collected by M.T., M.D., M.Y., Y.S. and H.Y. All authors contributed to the analyses and interpretation of the results.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Takeuchi, M., Doubell, M.J., Jackson, G.A. et al. Turbulence mediates marine aggregate formation and destruction in the upper ocean. Sci Rep 9, 16280 (2019). https://doi.org/10.1038/s41598-019-52470-5
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-019-52470-5
This article is cited by
-
A proteome scale study reveals how plastic surfaces and agitation promote protein aggregation
Scientific Reports (2023)
-
Persistent reshaping of cohesive sediment towards stable flocs by turbulence
Scientific Reports (2023)
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.