Vertical distribution of phytoplankton

Against this background, it seems appropriate to emphasise that the expectation of the vertical distributions of phytoplankton is that they should conform to the vertical differentiation of the water column, in terms of its current (or, at least, very recent) kinetic structure. The latter may comprise a wind-mixed convective layer overlying a typically less energetic layer of turbulence that is supposed to diminish with increasing depth, towards a benthic boundary layer in which turbulence is overcome by friction with the solid surface. However, in the seas as in lakes, the horizontal drift of the convective layer has to be compensated by counterflows, which movement promotes internal eddies and provides some turbulent kinetic energy from below. The formation of vertical density gradients may allow the confinement of the horizontal circulation to the upper part of the water column, leaving a distinct and kinetically rather inert water mass of the pycnocline, with very weak vertical motion. Density gradients form at depths in large, deep lakes and in the sea but do not contain basin-wide circulations; even though the gradients may persist, they may be rhythmically or chaotically displaced through the interplay of the gravitational 'sloshing' movements of deep water masses and the convective movements of the surface layer.

The vertical extent of the convective layer is, as we have seen, highly variable and subject to change at high frequency. It can vary from a few millimetres to tens of metres over a period of hours and between tens to hundreds of metres over a few days. The presence of density gradients reduces the entrainability of the deeper water into the surface flow. There is often great complexity at the interface (note, even if the top layer of the deeper water is sheared off and incorporated into the convective circulation of the upper, a gradient between the layers persists, albeit by now at a greater physical depth). Internal waves may form as a consequence of differential velocities, mirroring those formed at the water surface by the drag of high-velocity winds; and, just as surface waves break when the velocity differences can no longer be contained, so internal waves become unstable and collapse (Kelvin-Helmholtz instabilities: see, e.g., Imberger, 1985, for details), releasing more bottom water into the upper convection.

In shallow water columns, the extension of the convective layer is confined to the physical water depth, in which all energy-dissipative interactions and return flows must be accommodated. This necessarily results in very complex and aggressive mixing processes, often extending to the bottom boundary layer and, on occasions, penetrating it to the extent of entraining, and resuspending, the unconsolidated sedimented material.

Supposing wind-driven convective layers everywhere to be characterised by u* > 5 x 10-3ms-1 (i.e. 5mms-1), they should be capable of fulfilling the entrainment criterion for plankton with sinking, floating or swimming speeds of <250 x 10-6ms-1. Comparison with Section 2.6.1 supports the deduction that the distribution of almost all phytoplankters must be quickly randomised through the vertical extent of convective layers. However, if the driving energy weakens, so that the convective layer contracts (in line with the Monin-Obukhov prediction [Eq. 2.31] or, without simultaneous heat gain, because u* diminishes), plankters that were entrained towards the bottom of the layer become increasingly liable to disentrainment in situ, where, ergo, their own intrinsic movements begin to be expressed.

Four examples of distributional responses to physical structure are sufficient to demonstrate the basic behaviours of phytoplankters that are heavier than water (pc > pw), those that are frequently lighter than water (pc < pw), those non-motile species that are more nearly isopycnic (pc ~ Pw) and those that are sufficiently motile for any density difference to be at times surmountable. The selection also employs some of the

I Depth-time plot of the vertical distribution of Asterionella formosa in the North Basin of Windermere through 1947. Isopleths in live cells mL-1. The shaded area represents the extent of the metalimnion. Original from Lund et al. (1963), and redrawn from Reynolds (1984a).

I Depth-time plot of the vertical distribution of Asterionella formosa in the North Basin of Windermere through 1947. Isopleths in live cells mL-1. The shaded area represents the extent of the metalimnion. Original from Lund et al. (1963), and redrawn from Reynolds (1984a).

differing ways that distributional data may be represented.

Non-motile, negatively buoyant plankters (Pc > Pw)

For the first case, the classical study of Lund etal. (1963) on the season-long distribution of Asterionella cells in the North Basin of Windermere in 1947 is illustrated (Fig. 2.21). The densities of diatoms mostly exceed, sometimes considerably, that of the surrounding water in lakes and seas. Those bound to be negatively buoyant (they have positive sinking rates) are destined to be lost progressively from suspension, at variable rates that are due to the relationship between the (variable) intrinsic particle sinking rate and the (variable) depth of penetration of sufficient kinetic energy to fulfil the species-specific entrainment criterion. Even before the critical quantities were known, numerous studies had demonstrated the sensitivity of diatom distribution to water movements and to the onset of thermal stratification in particular, both in lakes (Ruttner, 1938; Findenegg, 1943; Lund, 1959; Nauwerck, 1963) and in the sea (Mar-galef, 1958, 1978; Parsons and Takahashi, 1973; Smayda, 1973,1980; Holligan and Harbour, 1977).

While numerous other factors intervene in the seasonal population dynamics and periodicity of diatoms, the importance of the depth of the surface mixed layer as another ecological threshold that must be satisfied for successful maintenance and recruitment of diatoms has been demonstrated many times (Reynolds and Wiseman, 1982; Reynolds etal., 1983b; Sommer, 1988a; Huisman et al., 1999; Padisak et al., 2003b). If the rate of recruitment through cell growth and replication fails to make good the aggregate rate of all losses, including to settlement, then the standing population must go into decline. The extent of the mixed depth is already implicated in the ability of a seed population to remain in suspension. As the Lycopodium experiments demonstrate (see Section 2.6.6), the nearer this contracts to the surface, the faster is the rate of loss from the diminishing mixed layer itself and from the enlarging, stagnating layer beneath it. This occurs independently of the chemical capacity of the water to support growth, although it is often influenced by spontaneous changes in the intrinsic sinking rate (Reynolds and Wiseman, 1982; Neale et al., 1991b).

The plot of Lund et al. (1963) (Fig. 2.21) shows the strong tendency in the first 3 months of the year towards vertical similarity in the concentration of Asterionella (with the isopleths, in cells mL-1, themselves vertically arranged), as their numbers slowly rise during the spring increase. With the progressive increase in day length and potential intensity of solar irradiance, the lake starts to stratify, with a pycnocline (represented in Fig. 2.21 by the fine stippling) developing at a depth of between 5 and 10 m from the surface. The contours reflect the segregating response of the vertical distribution, with an initial near-surface acceleration in recruitment but followed soon afterwards by rapid decline in numbers, as sinking losses by dilution from the truncated mixed layer overtake recruitment. The distribution of contours beneath the pycnocline acquire a diagonal trend (reflecting algal settlement) while the near-horizontal lines in the pycnocline itself confirm the heterogeneity of numbers in the vertical direction and the strong vertical gradient in algal concentration in the region of the pyc-nocline. It is not until the final breakdown of thermal stratification (usually in December in Windermere) that approximate homogeneity in the vertical is recovered.

Positively buoyant plankters (pc < pw)

The vertical distribution of buoyant organisms, which include many of the planktic, gas-vacuolate Cyanobacteria during at least stages of their development, is similarly responsive to variability in the diffusive strength of vertical convection, save that algae float, rather than sink, through the more stable layers. A further difference is that the population of the upper mixed layer potentially experiences concentration by net recruitment from upward-moving organisms rather than dilution as downward-moving organisms are shed from it.

It has long been appreciated that the formation of surface scums of buoyant Cyanobacteria (variously known, colloquially and in many languages, as water blooms, flowering of the waters, etc.: Reynolds and Walsby, 1975), and involving such genera as Anabaena, Anabaenopsis, Aphani-zomenon, Gloeotrichia, Gomphosphaeria, Woronichinia and (especially) Microcystis are prone to form in still, windless conditions (Griffiths, 1939). Buoyant Trichodesmium filaments also form locally dense surface patches in warm tropical seas under calm conditions (Ramamurthy, 1970): the little flakes of filaments also merited the sailors' colloquialism of 'sea sawdust'.

The mechanisms of scum formation are not straightforward but, rather, require the coincidence of three preconditions: a pre-existing population, a significant proportion of this being rendered positively buoyant on the balance of its gas-vesicle content, and the hydrographic conditions being such as to allow their disentrain-ment (Reynolds and Walsby, 1975). The present discussion assumes that the first two criteria are satisfied, scum formation now depending upon the relatively short-term onset of lowered diffu-sivity to the water surface, so that the magnitude of the flotation velocity (—ws) is no longer overwhelmed by the turbulent velocity (u*). The example presented in Fig. 2.22 traces the changing distribution of colonies of Microcystis aeruginosa, in relation to the thermal structure in a small temperate lake during one 24-h period of

Colonies / mL-1

0900 1700 2330 0730

Time of day/h

0900 1700 2330 0730

Time of day/h

1 Changes in the vertical distribution of Microcystis aeruginosa colonies in a small lake (shown as cylindrical curves) in relation to temperature (isopleths in °C), during 28/29 July, 1971 (SS, sunset; SR, sunrise). Data of Reynolds (1973b) and redrawn from Reynolds (1984a).

anticyclonic weather in July 1971. The colonies were, on average, buoyant throughout, having a mean flotation rate of ~9 ^m s-1 during the first day that increased almost twofold during the hours of darkness (Reynolds, 1973b). Note that an established temperature gradient, extending downwards from a depth of about 3.5 m, already contained the buoyant population and the changing vertical distribution of Microcystis occurred in relation to the secondary microstratification that developed during the course of the day (density gradient 0.1 kg m-3 m-1). A light breeze occurred in the early evening of 28 July, before windless conditions resumed. Some convectional cooling also occurred during the night, sufficient to redistribute the population to a small extent but not to dissipate the surface scum that had formed, which, from an average concentration of ~160 colonies mL-1, increased 37-fold (to 5940 mL-1 at 23.30).

The method of plotting, greatly favoured by the early plankton ecologists, invokes the use of 'cylindrical curves'. These are drawn as laterally viewed solid cones or more complex 'table-legs', the cross-sectional diameter at any given point being proportional to the cube root of the concentration. These shapes capture well the discontinuities in a given vertical distribution but may

0 0

Post a comment